diff --git a/.github/codeql/codeql-config.yml b/.github/codeql/codeql-config.yml index 9b6be27ab8..20807685e1 100644 --- a/.github/codeql/codeql-config.yml +++ b/.github/codeql/codeql-config.yml @@ -1,12 +1,19 @@ name: "LiteLLM CodeQL config" -# Exclude queries that produce result sets > 2 GiB on this codebase, -# causing 49+ minute runs that fail and block CI resources. +# Use security-extended suite instead of security-and-quality to avoid +# result sets > 2 GiB on this codebase that cause fatal OOM failures. +queries: + - uses: security-extended + +# These two queries are security queries included in security-extended that +# individually produce result sets > 2 GiB on this codebase, causing fatal +# OOM failures. Exclude them as a safety net until CI confirms they no longer +# OOM; drop these exclusions in a follow-up once verified. query-filters: - exclude: - id: py/clear-text-logging-sensitive-data # CWE-312/CleartextLogging.ql — result set > 2 GiB + id: py/clear-text-logging-sensitive-data # CWE-312 — > 2 GiB result set - exclude: - id: py/polynomial-redos # CWE-730/PolynomialReDoS.ql — result set > 2 GiB + id: py/polynomial-redos # CWE-730 — > 2 GiB result set paths-ignore: - tests diff --git a/docs/my-website/blog/gemini_embedding_2_multimodal/index.md b/docs/my-website/blog/gemini_embedding_2_multimodal/index.md new file mode 100644 index 0000000000..8c09432e3b --- /dev/null +++ b/docs/my-website/blog/gemini_embedding_2_multimodal/index.md @@ -0,0 +1,169 @@ +--- +slug: gemini_embedding_2_multimodal +title: "Gemini Embedding 2 Preview: Multimodal Embeddings on LiteLLM" +date: 2025-03-11T10:00:00 +authors: + - name: Sameer Kankute + title: SWE @ LiteLLM (LLM Translation) + url: https://www.linkedin.com/in/sameer-kankute/ + image_url: https://pbs.twimg.com/profile_images/2001352686994907136/ONgNuSk5_400x400.jpg +description: "Generate embeddings from text, images, audio, video, and PDFs with gemini-embedding-2-preview on LiteLLM via Gemini API and Vertex AI." +tags: [gemini, embeddings, multimodal, vertex ai] +hide_table_of_contents: false +--- + +import Tabs from '@theme/Tabs'; +import TabItem from '@theme/TabItem'; + +# Gemini Embedding 2 Preview: Multimodal Embeddings + +LiteLLM now supports **multimodal embeddings** with `gemini-embedding-2-preview`—generating a single embedding from a mix of text, images, audio, video, and PDF content. Available via both the **Gemini API** (API key) and **Vertex AI** (GCP credentials). + +## Supported Input Types + +| Modality | Supported Formats | +|----------|-------------------| +| **Text** | Plain text | +| **Image** | PNG, JPEG | +| **Audio** | MP3, WAV | +| **Video** | MP4, MOV | +| **Documents** | PDF | + +## Input Formats + +LiteLLM accepts three input formats for multimodal content: + +1. **Data URIs** – Base64-encoded inline: `data:image/png;base64,` +2. **GCS URLs** – Cloud Storage paths (Vertex AI): `gs://bucket/path/to/file.png` +3. **Gemini File References** – Pre-uploaded files (Gemini API): `files/abc123` + +## Quick Start + + + + +```python +from litellm import embedding +import os + +os.environ["GEMINI_API_KEY"] = "your-api-key" + +# Text + Image (base64) +response = embedding( + model="gemini/gemini-embedding-2-preview", + input=[ + "The food was delicious and the waiter...", + "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAIAQMAAAD+wSzIAAAABlBMVEX///+/v7+jQ3Y5AAAADklEQVQI12P4AIX8EAgALgAD/aNpbtEAAAAASUVORK5CYII" + ], +) +print(response) +``` + + + + + +```python +import litellm +from litellm import embedding + +litellm.vertex_project = "your-project-id" +litellm.vertex_location = "us-central1" + +# Text + Image (GCS URL) +response = embedding( + model="vertex_ai/gemini-embedding-2-preview", + input=[ + "Describe this image", + "gs://my-bucket/images/photo.png" + ], +) +print(response) +``` + + + + + +**1. Config (config.yaml)** + +```yaml +model_list: + - model_name: gemini-embedding-2-preview + litellm_params: + model: gemini/gemini-embedding-2-preview + api_key: os.environ/GEMINI_API_KEY + - model_name: vertex-gemini-embedding-2-preview + litellm_params: + model: vertex_ai/gemini-embedding-2-preview + vertex_project: os.environ/VERTEXAI_PROJECT + vertex_location: os.environ/VERTEXAI_LOCATION + +general_settings: + master_key: sk-1234 +``` + +**2. Start proxy** + +```bash +litellm --config config.yaml +``` + +**3. Call embeddings** + +```bash +curl -X POST http://localhost:4000/embeddings \ + -H "Authorization: Bearer sk-1234" \ + -H "Content-Type: application/json" \ + -d '{ + "model": "gemini-embedding-2-preview", + "input": [ + "The food was delicious and the waiter...", + "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAIAQMAAAD+wSzIAAAABlBMVEX///+/v7+jQ3Y5AAAADklEQVQI12P4AIX8EAgALgAD/aNpbtEAAAAASUVORK5CYII" + ] + }' +``` + + + + +## Input Format Examples + +| Format | Example | Provider | +|--------|---------|----------| +| **Data URI** | `data:image/png;base64,...` | Gemini, Vertex AI | +| **GCS URL** | `gs://bucket/path/image.png` | Vertex AI | +| **File reference** | `files/abc123` | Gemini API only | + +### Supported MIME Types for Data URIs + +- **Images:** `image/png`, `image/jpeg` +- **Audio:** `audio/mpeg`, `audio/wav` +- **Video:** `video/mp4`, `video/quicktime` +- **Documents:** `application/pdf` + +### GCS URL MIME Inference + +For Vertex AI, MIME types are inferred from file extensions: + +- `.png` → `image/png` +- `.jpg` / `.jpeg` → `image/jpeg` +- `.mp3` → `audio/mpeg` +- `.wav` → `audio/wav` +- `.mp4` → `video/mp4` +- `.mov` → `video/quicktime` +- `.pdf` → `application/pdf` + +## Optional Parameters + +| Parameter | Description | Maps to | +|-----------|-------------|---------| +| `dimensions` | Output embedding size | `outputDimensionality` | + +```python +response = embedding( + model="gemini/gemini-embedding-2-preview", + input=["text to embed"], + dimensions=768, # Optional: control output vector size +) +``` diff --git a/docs/my-website/docs/embedding/supported_embedding.md b/docs/my-website/docs/embedding/supported_embedding.md index 11ca4da48a..87acd0b33a 100644 --- a/docs/my-website/docs/embedding/supported_embedding.md +++ b/docs/my-website/docs/embedding/supported_embedding.md @@ -514,6 +514,57 @@ All models listed [here](https://ai.google.dev/gemini-api/docs/models/gemini) ar | Model Name | Function Call | | :--- | :--- | | text-embedding-004 | `embedding(model="gemini/text-embedding-004", input)` | +| gemini-embedding-2-preview | `embedding(model="gemini/gemini-embedding-2-preview", input)` | [Multimodal docs](#gemini-embedding-2-preview-multimodal) | + +### Gemini Embedding 2 Preview (Multimodal) + +`gemini-embedding-2-preview` supports **multimodal embeddings**—text, images, audio, video, and PDF in a single request. See [blog post](/blog/gemini_embedding_2_multimodal) for details. + +**Input formats:** +- **Data URIs:** `data:image/png;base64,` +- **Gemini file references:** `files/abc123` (pre-uploaded via Gemini Files API) + +**Supported MIME types:** `image/png`, `image/jpeg`, `audio/mpeg`, `audio/wav`, `video/mp4`, `video/quicktime`, `application/pdf` + + + + +```python +from litellm import embedding +import os +os.environ["GEMINI_API_KEY"] = "" + +# Text + Image (base64) +response = embedding( + model="gemini/gemini-embedding-2-preview", + input=[ + "The food was delicious and the waiter...", + "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAIAQMAAAD+wSzIAAAABlBMVEX///+/v7+jQ3Y5AAAADklEQVQI12P4AIX8EAgALgAD/aNpbtEAAAAASUVORK5CYII" + ], +) +print(response) +``` + + + + +```bash +curl -X POST http://localhost:4000/embeddings \ + -H "Authorization: Bearer sk-1234" \ + -H "Content-Type: application/json" \ + -d '{ + "model": "gemini-embedding-2-preview", + "input": [ + "The food was delicious and the waiter...", + "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAIAQMAAAD+wSzIAAAABlBMVEX///+/v7+jQ3Y5AAAADklEQVQI12P4AIX8EAgALgAD/aNpbtEAAAAASUVORK5CYII" + ] + }' +``` + + + + +**Optional:** `dimensions` maps to Gemini's `outputDimensionality`. ## Vertex AI Embedding Models diff --git a/docs/my-website/docs/mcp.md b/docs/my-website/docs/mcp.md index 600f69547d..b805cce4d7 100644 --- a/docs/my-website/docs/mcp.md +++ b/docs/my-website/docs/mcp.md @@ -133,6 +133,21 @@ LiteLLM attempts [OAuth 2.0 Authorization Server Discovery](https://datatracker.
+### AWS SigV4 Authentication + +For MCP servers hosted on [AWS Bedrock AgentCore](https://docs.aws.amazon.com/bedrock/latest/userguide/agentcore.html), select **AWS SigV4** as the authentication type. LiteLLM will sign every outgoing MCP request with your AWS credentials using [Signature Version 4](https://docs.aws.amazon.com/general/latest/gr/signature-version-4.html). + + + +Fill in your AWS region, service name (defaults to `bedrock-agentcore`), and optionally your AWS access key and secret. If credentials are omitted, LiteLLM falls back to the boto3 credential chain (IAM roles, environment variables, etc.). + +[**See full SigV4 setup guide**](./mcp_aws_sigv4.md) + +
+ ### Static Headers Sometimes your MCP server needs specific headers on every request. Maybe it's an API key, maybe it's a custom header the server expects. Instead of configuring auth, you can just set them directly. diff --git a/docs/my-website/docs/mcp_aws_sigv4.md b/docs/my-website/docs/mcp_aws_sigv4.md index e00cee4fd5..9dc60bce06 100644 --- a/docs/my-website/docs/mcp_aws_sigv4.md +++ b/docs/my-website/docs/mcp_aws_sigv4.md @@ -1,3 +1,7 @@ +import Tabs from '@theme/Tabs'; +import TabItem from '@theme/TabItem'; +import Image from '@theme/IdealImage'; + # MCP - AWS SigV4 Auth Use AWS SigV4 authentication to connect LiteLLM to MCP servers hosted on [AWS Bedrock AgentCore](https://docs.aws.amazon.com/bedrock/latest/userguide/agentcore.html). @@ -10,6 +14,36 @@ LiteLLM's `aws_sigv4` auth type handles this automatically: every outgoing MCP r ## Quick Start + + + +1. Navigate to **MCP Servers** and click **Add New MCP Server** +2. Set the transport to **Streamable HTTP** +3. Select **AWS SigV4** as the authentication type +4. Fill in your AWS credentials: + + + +
+ +| Field | Required | Description | +|-------|----------|-------------| +| **AWS Region** | Yes | AWS region for SigV4 signing (e.g., `us-east-1`) | +| **AWS Service Name** | No | Defaults to `bedrock-agentcore` | +| **AWS Access Key ID** | No | Falls back to boto3 credential chain if blank | +| **AWS Secret Access Key** | No | Required if Access Key ID is provided | +| **AWS Session Token** | No | Only needed for temporary STS credentials | + +Once created, LiteLLM will sign every outgoing MCP request with SigV4. The server's tools appear automatically in the MCP Tools list. + +**Editing credentials:** When editing an existing SigV4 server, leave credential fields blank to keep the current values. Only fields you fill in will be updated. + + + + ### 1. Set AWS credentials ```bash @@ -60,9 +94,12 @@ arn%3Aaws%3Abedrock-agentcore%3Aus-east-1%3A123456789012%3Aruntime%2Fmy-mcp-serv litellm --config config.yaml ``` -### 4. Use the MCP tools + + -Once started, your AgentCore MCP tools are available through LiteLLM like any other MCP server: +## Use the MCP tools + +Once configured, your AgentCore MCP tools are available through LiteLLM like any other MCP server: ```bash title="List available tools" curl http://localhost:4000/mcp-rest/tools/list \ diff --git a/docs/my-website/docs/providers/vertex_embedding.md b/docs/my-website/docs/providers/vertex_embedding.md index 5656ade337..9b530f2ae0 100644 --- a/docs/my-website/docs/providers/vertex_embedding.md +++ b/docs/my-website/docs/providers/vertex_embedding.md @@ -79,6 +79,7 @@ All models listed [here](https://github.com/BerriAI/litellm/blob/57f37f743886a02 | textembedding-gecko@003 | `embedding(model="vertex_ai/textembedding-gecko@003", input)` | | text-embedding-preview-0409 | `embedding(model="vertex_ai/text-embedding-preview-0409", input)` | | text-multilingual-embedding-preview-0409 | `embedding(model="vertex_ai/text-multilingual-embedding-preview-0409", input)` | +| gemini-embedding-2-preview | `embedding(model="vertex_ai/gemini-embedding-2-preview", input)` | [Multimodal docs](#gemini-embedding-2-preview-multimodal) | | Fine-tuned OR Custom Embedding models | `embedding(model="vertex_ai/", input)` | ### Supported OpenAI (Unified) Params @@ -257,6 +258,71 @@ model_list: ## **Multi-Modal Embeddings** +### Gemini Embedding 2 Preview (Multimodal) + +`gemini-embedding-2-preview` supports **unified multimodal embeddings**—text, images, audio, video, and PDF in a single request. See [blog post](/blog/gemini_embedding_2_multimodal) for details. + +**Input formats:** +- **Data URIs:** `data:image/png;base64,` +- **GCS URLs:** `gs://bucket/path/to/file.png` (MIME type inferred from extension) + +**Supported MIME types:** `image/png`, `image/jpeg`, `audio/mpeg`, `audio/wav`, `video/mp4`, `video/quicktime`, `application/pdf` + + + + +```python +import litellm +from litellm import embedding + +litellm.vertex_project = "your-project-id" +litellm.vertex_location = "us-central1" + +# Text + Image (GCS URL) +response = embedding( + model="vertex_ai/gemini-embedding-2-preview", + input=[ + "Describe this image", + "gs://my-bucket/images/photo.png" + ], +) + +# Text + Image (base64) +response = embedding( + model="vertex_ai/gemini-embedding-2-preview", + input=[ + "The food was delicious", + "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAIAQMAAAD+wSzIAAAABlBMVEX///+/v7+jQ3Y5AAAADklEQVQI12P4AIX8EAgALgAD/aNpbtEAAAAASUVORK5CYII" + ], +) +``` + + + + +```yaml +model_list: + - model_name: vertex-gemini-embedding-2-preview + litellm_params: + model: vertex_ai/gemini-embedding-2-preview + vertex_project: "your-project-id" + vertex_location: "us-central1" +``` + +```bash +curl -X POST http://localhost:4000/embeddings \ + -H "Authorization: Bearer sk-1234" \ + -H "Content-Type: application/json" \ + -d '{ + "model": "vertex-gemini-embedding-2-preview", + "input": ["Describe this", "gs://bucket/image.png"] + }' +``` + + + + +### multimodalembedding@001 (Legacy) Known Limitations: - Only supports 1 image / video / image per request diff --git a/docs/my-website/img/mcp_aws_sigv4_ui.png b/docs/my-website/img/mcp_aws_sigv4_ui.png new file mode 100644 index 0000000000..17016d3ae1 Binary files /dev/null and b/docs/my-website/img/mcp_aws_sigv4_ui.png differ diff --git a/litellm/llms/vertex_ai/common_utils.py b/litellm/llms/vertex_ai/common_utils.py index b9cbb52e57..078fce63cc 100644 --- a/litellm/llms/vertex_ai/common_utils.py +++ b/litellm/llms/vertex_ai/common_utils.py @@ -247,23 +247,27 @@ def _get_embedding_url( - bge/endpoint_id -> strips to endpoint_id for endpoints/ routing - numeric model -> routes to endpoints/ - regular model -> routes to publishers/google/models/ - """ - endpoint = "predict" - - # Strip routing prefixes (bge/, gemma/, etc.) for endpoint URL construction + - models with uses_embed_content flag -> use embedContent endpoint instead of predict + """ + original_model = model model = get_vertex_base_model_name(model=model) - # Get base URL (handles global vs regional) + try: + model_info = litellm.get_model_info( + model=original_model, + custom_llm_provider="vertex_ai", + ) + uses_embed_content = model_info.get("uses_embed_content", False) + except Exception: + uses_embed_content = False + + endpoint = "embedContent" if uses_embed_content else "predict" + base_url = get_vertex_base_url(vertex_location) if model.isdigit(): - # https://us-central1-aiplatform.googleapis.com/v1/projects/$PROJECT_ID/locations/us-central1/endpoints/$ENDPOINT_ID:predict - # https://aiplatform.googleapis.com/v1/projects/$PROJECT_ID/locations/global/endpoints/$ENDPOINT_ID:predict url = f"{base_url}/{vertex_api_version}/projects/{vertex_project}/locations/{vertex_location}/endpoints/{model}:{endpoint}" else: - # Regular model -> publisher model - # https://us-central1-aiplatform.googleapis.com/v1/projects/$PROJECT_ID/locations/us-central1/publishers/google/models/{model}:predict - # https://aiplatform.googleapis.com/v1/projects/$PROJECT_ID/locations/global/publishers/google/models/{model}:predict url = f"{base_url}/v1/projects/{vertex_project}/locations/{vertex_location}/publishers/google/models/{model}:{endpoint}" return url, endpoint diff --git a/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_handler.py b/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_handler.py index 07f57a4a7f..68901340c7 100644 --- a/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_handler.py +++ b/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_handler.py @@ -3,12 +3,11 @@ Google AI Studio /batchEmbedContents Embeddings Endpoint """ import json -from typing import Any, Literal, Optional, Union +from typing import Any, Dict, Literal, Optional, Union import httpx import litellm -from litellm.types.utils import EmbeddingResponse from litellm.llms.custom_httpx.http_handler import ( AsyncHTTPHandler, HTTPHandler, @@ -19,15 +18,98 @@ from litellm.types.llms.vertex_ai import ( VertexAIBatchEmbeddingsRequestBody, VertexAIBatchEmbeddingsResponseObject, ) +from litellm.types.utils import EmbeddingResponse from ..gemini.vertex_and_google_ai_studio_gemini import VertexLLM from .batch_embed_content_transformation import ( + _is_file_reference, + _is_multimodal_input, + process_embed_content_response, process_response, transform_openai_input_gemini_content, + transform_openai_input_gemini_embed_content, ) class GoogleBatchEmbeddings(VertexLLM): + def _resolve_file_references( + self, + input: EmbeddingInput, + api_key: str, + sync_handler: HTTPHandler, + ) -> Dict[str, Dict[str, str]]: + """ + Resolve Gemini file references (files/...) to get mime_type and uri. + + Args: + input: EmbeddingInput that may contain file references + api_key: Gemini API key + sync_handler: HTTP client + + Returns: + Dict mapping file name to {mime_type, uri} + """ + input_list = [input] if isinstance(input, str) else input + resolved_files: Dict[str, Dict[str, str]] = {} + + for element in input_list: + if isinstance(element, str) and _is_file_reference(element): + url = f"https://generativelanguage.googleapis.com/v1beta/{element}" + headers = {"x-goog-api-key": api_key} + response = sync_handler.get(url=url, headers=headers) + + if response.status_code != 200: + raise Exception( + f"Error fetching file {element}: {response.status_code} {response.text}" + ) + + file_data = response.json() + resolved_files[element] = { + "mime_type": file_data.get("mimeType", ""), + "uri": file_data.get("uri", element), + } + + return resolved_files + + async def _async_resolve_file_references( + self, + input: EmbeddingInput, + api_key: str, + async_handler: AsyncHTTPHandler, + ) -> Dict[str, Dict[str, str]]: + """ + Async version of _resolve_file_references. + + Args: + input: EmbeddingInput that may contain file references + api_key: Gemini API key + async_handler: Async HTTP client + + Returns: + Dict mapping file name to {mime_type, uri} + """ + input_list = [input] if isinstance(input, str) else input + resolved_files: Dict[str, Dict[str, str]] = {} + + for element in input_list: + if isinstance(element, str) and _is_file_reference(element): + url = f"https://generativelanguage.googleapis.com/v1beta/{element}" + headers = {"x-goog-api-key": api_key} + response = await async_handler.get(url=url, headers=headers) + + if response.status_code != 200: + raise Exception( + f"Error fetching file {element}: {response.status_code} {response.text}" + ) + + file_data = response.json() + resolved_files[element] = { + "mime_type": file_data.get("mimeType", ""), + "uri": file_data.get("uri", element), + } + + return resolved_files + def batch_embeddings( self, model: str, @@ -54,20 +136,6 @@ class GoogleBatchEmbeddings(VertexLLM): custom_llm_provider=custom_llm_provider, ) - auth_header, url = self._get_token_and_url( - model=model, - auth_header=_auth_header, - gemini_api_key=api_key, - vertex_project=vertex_project, - vertex_location=vertex_location, - vertex_credentials=vertex_credentials, - stream=None, - custom_llm_provider=custom_llm_provider, - api_base=api_base, - should_use_v1beta1_features=False, - mode="batch_embedding", - ) - if client is None: _params = {} if timeout is not None: @@ -83,9 +151,25 @@ class GoogleBatchEmbeddings(VertexLLM): optional_params = optional_params or {} - ### TRANSFORMATION ### - request_data = transform_openai_input_gemini_content( - input=input, model=model, optional_params=optional_params + is_multimodal = _is_multimodal_input(input) + use_embed_content = is_multimodal or (custom_llm_provider == "vertex_ai") + if use_embed_content: + mode = "embedding" + else: + mode = "batch_embedding" + + auth_header, url = self._get_token_and_url( + model=model, + auth_header=_auth_header, + gemini_api_key=api_key, + vertex_project=vertex_project, + vertex_location=vertex_location, + vertex_credentials=vertex_credentials, + stream=None, + custom_llm_provider=custom_llm_provider, + api_base=api_base, + should_use_v1beta1_features=False, + mode=mode, ) headers = { @@ -93,14 +177,46 @@ class GoogleBatchEmbeddings(VertexLLM): } if auth_header is not None: if isinstance(auth_header, dict): - # For Gemini with custom api_base: auth_header is {"x-goog-api-key": "..."} headers.update(auth_header) else: - # For Vertex AI: auth_header is a Bearer token string headers["Authorization"] = f"Bearer {auth_header}" if extra_headers is not None: headers.update(extra_headers) + if aembedding is True: + return self.async_batch_embeddings( # type: ignore + model=model, + api_base=api_base, + url=url, + data=None, + model_response=model_response, + timeout=timeout, + headers=headers, + input=input, + use_embed_content=use_embed_content, + api_key=api_key, + optional_params=optional_params, + logging_obj=logging_obj, + ) + + ### TRANSFORMATION (sync path) ### + if use_embed_content: + resolved_files = {} + if api_key: + resolved_files = self._resolve_file_references( + input=input, api_key=api_key, sync_handler=sync_handler + ) + request_data = transform_openai_input_gemini_embed_content( + input=input, + model=model, + optional_params=optional_params, + resolved_files=resolved_files, + ) + else: + request_data = transform_openai_input_gemini_content( + input=input, model=model, optional_params=optional_params + ) + ## LOGGING logging_obj.pre_call( input=input, @@ -112,18 +228,6 @@ class GoogleBatchEmbeddings(VertexLLM): }, ) - if aembedding is True: - return self.async_batch_embeddings( # type: ignore - model=model, - api_base=api_base, - url=url, - data=request_data, - model_response=model_response, - timeout=timeout, - headers=headers, - input=input, - ) - response = sync_handler.post( url=url, headers=headers, @@ -134,26 +238,38 @@ class GoogleBatchEmbeddings(VertexLLM): raise Exception(f"Error: {response.status_code} {response.text}") _json_response = response.json() - _predictions = VertexAIBatchEmbeddingsResponseObject(**_json_response) # type: ignore - - return process_response( - model=model, - model_response=model_response, - _predictions=_predictions, - input=input, - ) + + if use_embed_content: + return process_embed_content_response( + input=input, + model_response=model_response, + model=model, + response_json=_json_response, + ) + else: + _predictions = VertexAIBatchEmbeddingsResponseObject(**_json_response) # type: ignore + return process_response( + model=model, + model_response=model_response, + _predictions=_predictions, + input=input, + ) async def async_batch_embeddings( self, model: str, api_base: Optional[str], url: str, - data: VertexAIBatchEmbeddingsRequestBody, + data: Optional[Union[VertexAIBatchEmbeddingsRequestBody, dict]], model_response: EmbeddingResponse, input: EmbeddingInput, timeout: Optional[Union[float, httpx.Timeout]], headers={}, client: Optional[AsyncHTTPHandler] = None, + use_embed_content: bool = False, + api_key: Optional[str] = None, + optional_params: Optional[dict] = None, + logging_obj: Optional[Any] = None, ) -> EmbeddingResponse: if client is None: _params = {} @@ -171,6 +287,36 @@ class GoogleBatchEmbeddings(VertexLLM): else: async_handler = client # type: ignore + ### TRANSFORMATION (async path) ### + if use_embed_content: + resolved_files = {} + if api_key: + resolved_files = await self._async_resolve_file_references( + input=input, api_key=api_key, async_handler=async_handler + ) + data = transform_openai_input_gemini_embed_content( + input=input, + model=model, + optional_params=optional_params or {}, + resolved_files=resolved_files, + ) + else: + data = transform_openai_input_gemini_content( + input=input, model=model, optional_params=optional_params or {} + ) + + ## LOGGING + if logging_obj is not None: + logging_obj.pre_call( + input=input, + api_key="", + additional_args={ + "complete_input_dict": data, + "api_base": url, + "headers": headers, + }, + ) + response = await async_handler.post( url=url, headers=headers, @@ -181,11 +327,19 @@ class GoogleBatchEmbeddings(VertexLLM): raise Exception(f"Error: {response.status_code} {response.text}") _json_response = response.json() - _predictions = VertexAIBatchEmbeddingsResponseObject(**_json_response) # type: ignore - - return process_response( - model=model, - model_response=model_response, - _predictions=_predictions, - input=input, - ) + + if use_embed_content: + return process_embed_content_response( + input=input, + model_response=model_response, + model=model, + response_json=_json_response, + ) + else: + _predictions = VertexAIBatchEmbeddingsResponseObject(**_json_response) # type: ignore + return process_response( + model=model, + model_response=model_response, + _predictions=_predictions, + input=input, + ) diff --git a/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_transformation.py b/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_transformation.py index 455ec1d18f..41f477d9db 100644 --- a/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_transformation.py +++ b/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_transformation.py @@ -4,20 +4,142 @@ Transformation logic from OpenAI /v1/embeddings format to Google AI Studio /batc Why separate file? Make it easy to see how transformation works """ -from typing import List +from typing import Dict, List, Optional, Tuple -from litellm.types.utils import EmbeddingResponse from litellm.types.llms.openai import EmbeddingInput from litellm.types.llms.vertex_ai import ( + BlobType, ContentType, EmbedContentRequest, + FileDataType, PartType, VertexAIBatchEmbeddingsRequestBody, VertexAIBatchEmbeddingsResponseObject, ) -from litellm.types.utils import Embedding, Usage +from litellm.types.utils import Embedding, EmbeddingResponse, Usage from litellm.utils import get_formatted_prompt, token_counter +SUPPORTED_EMBEDDING_MIME_TYPES = { + "image/png", + "image/jpeg", + "audio/mpeg", + "audio/wav", + "video/mp4", + "video/quicktime", + "application/pdf", +} + + +def _is_file_reference(s: str) -> bool: + """Check if string is a Gemini file reference (files/...).""" + return isinstance(s, str) and s.startswith("files/") + + +def _is_gcs_url(s: str) -> bool: + """Check if string is a GCS URL (gs://...).""" + return isinstance(s, str) and s.startswith("gs://") + + +def _infer_mime_type_from_gcs_url(gcs_url: str) -> str: + """ + Infer MIME type from GCS URL file extension. + + Args: + gcs_url: GCS URL like gs://bucket/path/to/file.png + + Returns: + str: Inferred MIME type + + Raises: + ValueError: If file extension is not supported + """ + extension_to_mime = { + ".png": "image/png", + ".jpg": "image/jpeg", + ".jpeg": "image/jpeg", + ".mp3": "audio/mpeg", + ".wav": "audio/wav", + ".mp4": "video/mp4", + ".mov": "video/quicktime", + ".pdf": "application/pdf", + } + + gcs_url_lower = gcs_url.lower() + for ext, mime_type in extension_to_mime.items(): + if gcs_url_lower.endswith(ext): + return mime_type + + raise ValueError( + f"Unable to infer MIME type from GCS URL: {gcs_url}. " + f"Supported extensions: {', '.join(extension_to_mime.keys())}" + ) + + +def _parse_data_url(data_url: str) -> Tuple[str, str]: + """ + Parse a data URL to extract the media type and base64 data. + + Args: + data_url: Data URL in format: data:image/jpeg;base64,/9j/4AAQ... + + Returns: + tuple: (media_type, base64_data) + media_type: e.g., "image/jpeg", "video/mp4", "audio/mpeg" + base64_data: The base64-encoded data without the prefix + + Raises: + ValueError: If data URL format is invalid or MIME type is unsupported + """ + if not data_url.startswith("data:"): + raise ValueError(f"Invalid data URL format: {data_url[:50]}...") + + if "," not in data_url: + raise ValueError(f"Invalid data URL format (missing comma): {data_url[:50]}...") + + metadata, base64_data = data_url.split(",", 1) + + metadata = metadata[5:] + + if ";" in metadata: + media_type = metadata.split(";")[0] + else: + media_type = metadata + + if media_type not in SUPPORTED_EMBEDDING_MIME_TYPES: + raise ValueError( + f"Unsupported MIME type for embedding: {media_type}. " + f"Supported types: {', '.join(sorted(SUPPORTED_EMBEDDING_MIME_TYPES))}" + ) + + return media_type, base64_data + + +def _is_multimodal_input(input: EmbeddingInput) -> bool: + """ + Check if the input contains multimodal data (data URIs, file references, or GCS URLs). + + Args: + input: EmbeddingInput (str or List[str]) + + Returns: + bool: True if any element is a data URI, file reference, or GCS URL + """ + if isinstance(input, str): + input_list = [input] + else: + input_list = input + + for element in input_list: + if isinstance(element, str): + if element.startswith("data:") and ";base64," in element: + return True + if _is_file_reference(element): + return True + if _is_gcs_url(element): + return True + + return False + def transform_openai_input_gemini_content( input: EmbeddingInput, model: str, optional_params: dict @@ -26,12 +148,17 @@ def transform_openai_input_gemini_content( The content to embed. Only the parts.text fields will be counted. """ gemini_model_name = "models/{}".format(model) + + gemini_params = optional_params.copy() + if "dimensions" in gemini_params: + gemini_params["outputDimensionality"] = gemini_params.pop("dimensions") + requests: List[EmbedContentRequest] = [] if isinstance(input, str): request = EmbedContentRequest( model=gemini_model_name, content=ContentType(parts=[PartType(text=input)]), - **optional_params + **gemini_params ) requests.append(request) else: @@ -39,13 +166,119 @@ def transform_openai_input_gemini_content( request = EmbedContentRequest( model=gemini_model_name, content=ContentType(parts=[PartType(text=i)]), - **optional_params + **gemini_params ) requests.append(request) return VertexAIBatchEmbeddingsRequestBody(requests=requests) +def transform_openai_input_gemini_embed_content( + input: EmbeddingInput, + model: str, + optional_params: dict, + resolved_files: Optional[Dict[str, Dict[str, str]]] = None, +) -> dict: + """ + Transform OpenAI embedding input to Gemini embedContent format (multimodal). + + Args: + input: EmbeddingInput (str or List[str]) with text, data URIs, or file references + model: Model name + optional_params: Additional parameters (taskType, outputDimensionality, etc.) + resolved_files: Dict mapping file names (files/abc) to {mime_type, uri} + + Returns: + dict: Gemini embedContent request body with content.parts + """ + resolved_files = resolved_files or {} + + gemini_params = optional_params.copy() + if "dimensions" in gemini_params: + gemini_params["outputDimensionality"] = gemini_params.pop("dimensions") + + input_list = [input] if isinstance(input, str) else input + parts: List[PartType] = [] + + for element in input_list: + if not isinstance(element, str): + raise ValueError(f"Unsupported input type: {type(element)}") + + if element.startswith("data:") and ";base64," in element: + mime_type, base64_data = _parse_data_url(element) + blob: BlobType = {"mime_type": mime_type, "data": base64_data} + parts.append(PartType(inline_data=blob)) + elif _is_gcs_url(element): + mime_type = _infer_mime_type_from_gcs_url(element) + file_data: FileDataType = { + "mime_type": mime_type, + "file_uri": element, + } + parts.append(PartType(file_data=file_data)) + elif _is_file_reference(element): + if element not in resolved_files: + raise ValueError(f"File reference {element} not resolved") + file_info = resolved_files[element] + file_data_ref: FileDataType = { + "mime_type": file_info["mime_type"], + "file_uri": file_info["uri"], + } + parts.append(PartType(file_data=file_data_ref)) + else: + parts.append(PartType(text=element)) + + request_body: dict = { + "content": ContentType(parts=parts), + **gemini_params, + } + + return request_body + + +def process_embed_content_response( + input: EmbeddingInput, + model_response: EmbeddingResponse, + model: str, + response_json: dict, +) -> EmbeddingResponse: + """ + Process Gemini embedContent response (single embedding for multimodal input). + + Args: + input: Original input + model_response: EmbeddingResponse to populate + model: Model name + response_json: Raw JSON response from embedContent endpoint + + Returns: + EmbeddingResponse with single embedding + """ + if "embedding" not in response_json: + raise ValueError(f"embedContent response missing 'embedding' field: {response_json}") + + embedding_data = response_json["embedding"] + + openai_embedding = Embedding( + embedding=embedding_data["values"], + index=0, + object="embedding", + ) + + model_response.data = [openai_embedding] + model_response.model = model + + if _is_multimodal_input(input): + prompt_tokens = 0 + else: + input_text = get_formatted_prompt(data={"input": input}, call_type="embedding") + prompt_tokens = token_counter(model=model, text=input_text) + model_response.usage = Usage( + prompt_tokens=prompt_tokens, total_tokens=prompt_tokens + ) + + return model_response + + def process_response( input: EmbeddingInput, model_response: EmbeddingResponse, diff --git a/litellm/main.py b/litellm/main.py index 40f3638496..2b210c79a5 100644 --- a/litellm/main.py +++ b/litellm/main.py @@ -132,6 +132,7 @@ from litellm.utils import ( create_tokenizer, get_api_key, get_llm_provider, + get_model_info, get_non_default_completion_params, get_non_default_transcription_params, get_optional_params_embeddings, @@ -5190,13 +5191,37 @@ def embedding( # noqa: PLR0915 or get_secret_str("VERTEX_API_BASE") ) - if ( + try: + model_info = get_model_info(model=model, custom_llm_provider="vertex_ai") + uses_embed_content = model_info.get("uses_embed_content", False) + except Exception: + uses_embed_content = False + + if uses_embed_content: + response = google_batch_embeddings.batch_embeddings( # type: ignore + model=model, + input=input, + encoding=_get_encoding(), + logging_obj=logging, + optional_params=optional_params, + model_response=EmbeddingResponse(), + vertex_project=vertex_ai_project, + vertex_location=vertex_ai_location, + vertex_credentials=vertex_credentials, + aembedding=aembedding, + print_verbose=print_verbose, + custom_llm_provider="vertex_ai", + api_key=None, + api_base=api_base, + client=client, + extra_headers=headers, + ) + elif ( "image" in optional_params or "video" in optional_params or model in vertex_multimodal_embedding.SUPPORTED_MULTIMODAL_EMBEDDING_MODELS ): - # multimodal embedding is supported on vertex httpx response = vertex_multimodal_embedding.multimodal_embedding( model=model, input=input, diff --git a/litellm/model_prices_and_context_window_backup.json b/litellm/model_prices_and_context_window_backup.json index b8de3f04df..d9ba7eb0cc 100644 --- a/litellm/model_prices_and_context_window_backup.json +++ b/litellm/model_prices_and_context_window_backup.json @@ -2565,6 +2565,19 @@ "supports_parallel_function_calling": true, "supports_tool_choice": true }, + "azure/gpt-35-turbo-1106": { + "deprecation_date": "2025-03-31", + "input_cost_per_token": 1e-06, + "litellm_provider": "azure", + "max_input_tokens": 16384, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_token": 2e-06, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true + }, "azure/gpt-35-turbo-16k": { "input_cost_per_token": 3e-06, "litellm_provider": "azure", @@ -8045,6 +8058,20 @@ "supports_reasoning": true, "supports_tool_choice": true }, + "cerebras/zai-glm-4.6": { + "deprecation_date": "2026-01-20", + "input_cost_per_token": 2.25e-06, + "litellm_provider": "cerebras", + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 2.75e-06, + "source": "https://www.cerebras.ai/pricing", + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, "cerebras/zai-glm-4.7": { "input_cost_per_token": 2.25e-06, "litellm_provider": "cerebras", @@ -13697,6 +13724,96 @@ "supports_vision": true, "supports_web_search": true }, + "gemini-live-2.5-flash-preview-native-audio-09-2025": { + "cache_read_input_token_cost": 7.5e-08, + "input_cost_per_audio_token": 3e-06, + "input_cost_per_token": 3e-07, + "litellm_provider": "vertex_ai-language-models", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "realtime", + "output_cost_per_audio_token": 1.2e-05, + "output_cost_per_token": 2e-06, + "source": "https://ai.google.dev/gemini-api/docs/pricing", + "supported_endpoints": [ + "/vertex_ai/live" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true + }, + "gemini/gemini-live-2.5-flash-preview-native-audio-09-2025": { + "cache_read_input_token_cost": 7.5e-08, + "input_cost_per_audio_token": 3e-06, + "input_cost_per_token": 3e-07, + "litellm_provider": "gemini", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "realtime", + "output_cost_per_audio_token": 1.2e-05, + "output_cost_per_token": 2e-06, + "rpm": 100000, + "source": "https://ai.google.dev/gemini-api/docs/pricing", + "supported_endpoints": [ + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true, + "tpm": 8000000 + }, "gemini-2.5-flash-lite-preview-06-17": { "deprecation_date": "2025-11-18", "cache_read_input_token_cost": 2.5e-08, @@ -14321,6 +14438,32 @@ "output_vector_size": 3072, "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models" }, + "gemini-embedding-2-preview": { + "input_cost_per_audio_per_second": 0.00016, + "input_cost_per_image": 0.00012, + "input_cost_per_token": 2e-07, + "input_cost_per_video_per_second": 0.0237, + "litellm_provider": "vertex_ai-embedding-models", + "max_input_tokens": 8192, + "max_tokens": 8192, + "mode": "embedding", + "output_cost_per_token": 0, + "output_vector_size": 3072, + "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing", + "uses_embed_content": true + }, + "vertex_ai/gemini-embedding-2-preview": { + "input_cost_per_token": 1.5e-07, + "litellm_provider": "vertex_ai", + "max_input_tokens": 8192, + "max_tokens": 8192, + "mode": "embedding", + "output_cost_per_token": 0, + "output_vector_size": 3072, + "source": "https://ai.google.dev/gemini-api/docs/embeddings#multimodal", + "supports_multimodal": true, + "uses_embed_content": true + }, "gemini/gemini-embedding-001": { "input_cost_per_token": 1.5e-07, "litellm_provider": "gemini", @@ -14333,6 +14476,19 @@ "source": "https://ai.google.dev/gemini-api/docs/embeddings#model-versions", "tpm": 10000000 }, + "gemini/gemini-embedding-2-preview": { + "input_cost_per_token": 1.5e-07, + "litellm_provider": "gemini", + "max_input_tokens": 8192, + "max_tokens": 8192, + "mode": "embedding", + "output_cost_per_token": 0, + "output_vector_size": 3072, + "rpm": 10000, + "source": "https://ai.google.dev/gemini-api/docs/embeddings#multimodal", + "supports_multimodal": true, + "tpm": 10000000 + }, "gemini/gemini-2.0-flash": { "cache_read_input_token_cost": 2.5e-08, "deprecation_date": "2026-06-01", @@ -16588,6 +16744,20 @@ "supports_system_messages": true, "supports_tool_choice": true }, + "gpt-4-0613": { + "deprecation_date": "2025-06-06", + "input_cost_per_token": 3e-05, + "litellm_provider": "openai", + "max_input_tokens": 8192, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_token": 6e-05, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, "gpt-4-1106-preview": { "deprecation_date": "2026-03-26", "input_cost_per_token": 1e-05, @@ -27107,6 +27277,16 @@ "output_cost_per_token": 0.0, "output_vector_size": 1536 }, + "text-embedding-ada-002-v2": { + "input_cost_per_token": 1e-07, + "input_cost_per_token_batches": 5e-08, + "litellm_provider": "openai", + "max_input_tokens": 8191, + "max_tokens": 8191, + "mode": "embedding", + "output_cost_per_token": 0.0, + "output_cost_per_token_batches": 0.0 + }, "text-embedding-large-exp-03-07": { "input_cost_per_character": 2.5e-08, "input_cost_per_token": 1e-07, @@ -36686,32 +36866,5 @@ "supports_reasoning": true, "supports_response_schema": true, "supports_tool_choice": true - }, - "azure/gpt-35-turbo-1106": { - "deprecation_date": "2025-03-31", - "input_cost_per_token": 1e-06, - "litellm_provider": "azure", - "max_input_tokens": 16384, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_token": 2e-06, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true - }, - "cerebras/zai-glm-4.6": { - "deprecation_date": "2026-01-20", - "input_cost_per_token": 2.25e-06, - "litellm_provider": "cerebras", - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 2.75e-06, - "source": "https://www.cerebras.ai/pricing", - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true } } diff --git a/litellm/proxy/_experimental/mcp_server/db.py b/litellm/proxy/_experimental/mcp_server/db.py index 93580b5430..c6af4be314 100644 --- a/litellm/proxy/_experimental/mcp_server/db.py +++ b/litellm/proxy/_experimental/mcp_server/db.py @@ -100,6 +100,50 @@ def encrypt_credentials( value=client_secret, new_encryption_key=encryption_key, ) + # AWS SigV4 credential fields + aws_access_key_id = credentials.get("aws_access_key_id") + if aws_access_key_id is not None: + credentials["aws_access_key_id"] = encrypt_value_helper( + value=aws_access_key_id, + new_encryption_key=encryption_key, + ) + aws_secret_access_key = credentials.get("aws_secret_access_key") + if aws_secret_access_key is not None: + credentials["aws_secret_access_key"] = encrypt_value_helper( + value=aws_secret_access_key, + new_encryption_key=encryption_key, + ) + aws_session_token = credentials.get("aws_session_token") + if aws_session_token is not None: + credentials["aws_session_token"] = encrypt_value_helper( + value=aws_session_token, + new_encryption_key=encryption_key, + ) + # aws_region_name and aws_service_name are NOT secrets — stored as-is + return credentials + + +def decrypt_credentials( + credentials: MCPCredentials, +) -> MCPCredentials: + """Decrypt all secret fields in an MCPCredentials dict using the global salt key.""" + secret_fields = [ + "auth_value", + "client_id", + "client_secret", + "aws_access_key_id", + "aws_secret_access_key", + "aws_session_token", + ] + for field in secret_fields: + value = credentials.get(field) + if value is not None: + credentials[field] = decrypt_value_helper( + value=value, + key=field, + exception_type="debug", + return_original_value=True, + ) return credentials @@ -350,9 +394,57 @@ async def update_mcp_server( """ Update a new mcp server record in the db """ + import json + + from litellm.litellm_core_utils.safe_json_dumps import safe_dumps + # Use helper to prepare data with proper JSON serialization data_dict = _prepare_mcp_server_data(data) + # Pre-fetch existing record once if we need it for auth_type or credential logic + existing = None + has_credentials = "credentials" in data_dict and data_dict["credentials"] is not None + if data.auth_type or has_credentials: + existing = await prisma_client.db.litellm_mcpservertable.find_unique( + where={"server_id": data.server_id} + ) + + # Clear stale credentials when auth_type changes but no new credentials provided + if ( + data.auth_type + and "credentials" not in data_dict + and existing + and existing.auth_type is not None + and existing.auth_type != data.auth_type + ): + data_dict["credentials"] = None + + # Merge credentials: preserve existing fields not present in the update. + # Without this, a partial credential update (e.g. changing only region) + # would wipe encrypted secrets that the UI cannot display back. + if "credentials" in data_dict and data_dict["credentials"] is not None: + if existing and existing.credentials: + # Only merge when auth_type is unchanged. Switching auth types + # (e.g. oauth2 → api_key) should replace credentials entirely + # to avoid stale secrets from the previous auth type lingering. + auth_type_unchanged = ( + data.auth_type is None or data.auth_type == existing.auth_type + ) + if auth_type_unchanged: + existing_creds = ( + json.loads(existing.credentials) + if isinstance(existing.credentials, str) + else dict(existing.credentials) + ) + new_creds = ( + json.loads(data_dict["credentials"]) + if isinstance(data_dict["credentials"], str) + else dict(data_dict["credentials"]) + ) + # New values override existing; existing keys not in update are preserved + merged = {**existing_creds, **new_creds} + data_dict["credentials"] = safe_dumps(merged) + # Add audit fields data_dict["updated_by"] = touched_by @@ -374,8 +466,12 @@ async def rotate_mcp_server_credentials_master_key( continue credentials_copy = dict(credentials) - encrypted_credentials = encrypt_credentials( + # Decrypt with current key first, then re-encrypt with new key + decrypted_credentials = decrypt_credentials( credentials=cast(MCPCredentials, credentials_copy), + ) + encrypted_credentials = encrypt_credentials( + credentials=decrypted_credentials, encryption_key=new_master_key, ) diff --git a/litellm/proxy/_experimental/mcp_server/mcp_server_manager.py b/litellm/proxy/_experimental/mcp_server/mcp_server_manager.py index 8769d3952a..b10bfde491 100644 --- a/litellm/proxy/_experimental/mcp_server/mcp_server_manager.py +++ b/litellm/proxy/_experimental/mcp_server/mcp_server_manager.py @@ -597,9 +597,10 @@ class MCPServerManager: else: client_secret_value = encrypted_client_secret - # TODO: Add AWS SigV4 credential decryption here when DB-stored - # SigV4 MCP servers are supported. Requires corresponding changes - # to encrypt_credentials() in db.py and MCPCredentials TypedDict. + # AWS SigV4 credential fields + aws_creds = self._extract_aws_credentials( + credentials_dict, credentials_are_encrypted + ) scopes: Optional[List[str]] = None if credentials_dict: @@ -679,6 +680,12 @@ class MCPServerManager: is_byok=bool(getattr(mcp_server, "is_byok", False)), byok_description=getattr(mcp_server, "byok_description", None) or [], byok_api_key_help_url=getattr(mcp_server, "byok_api_key_help_url", None), + # AWS SigV4 fields + aws_access_key_id=aws_creds.get("aws_access_key_id"), + aws_secret_access_key=aws_creds.get("aws_secret_access_key"), + aws_session_token=aws_creds.get("aws_session_token"), + aws_region_name=aws_creds.get("aws_region_name"), + aws_service_name=aws_creds.get("aws_service_name"), ) return new_server @@ -1518,6 +1525,52 @@ class MCPServerManager: return None + @staticmethod + def _decrypt_credential_field( + encrypted_value: Optional[str], + key: str, + credentials_are_encrypted: bool, + ) -> Optional[str]: + """Decrypt a single credential field, or return as-is if not encrypted.""" + if not encrypted_value: + return None + if credentials_are_encrypted: + return decrypt_value_helper( + value=encrypted_value, + key=key, + exception_type="debug", + return_original_value=True, + ) + return encrypted_value + + def _extract_aws_credentials( + self, + credentials_dict: Optional[Dict[str, str]], + credentials_are_encrypted: bool, + ) -> Dict[str, Optional[str]]: + """Extract and decrypt AWS SigV4 credential fields from credentials dict.""" + if not credentials_dict: + return {} + return { + "aws_access_key_id": self._decrypt_credential_field( + credentials_dict.get("aws_access_key_id"), + "aws_access_key_id", + credentials_are_encrypted, + ), + "aws_secret_access_key": self._decrypt_credential_field( + credentials_dict.get("aws_secret_access_key"), + "aws_secret_access_key", + credentials_are_encrypted, + ), + "aws_session_token": self._decrypt_credential_field( + credentials_dict.get("aws_session_token"), + "aws_session_token", + credentials_are_encrypted, + ), + "aws_region_name": credentials_dict.get("aws_region_name"), + "aws_service_name": credentials_dict.get("aws_service_name"), + } + def _extract_scopes(self, scopes_value: Any) -> Optional[List[str]]: if isinstance(scopes_value, str): scopes = [s.strip() for s in scopes_value.split() if s.strip()] diff --git a/litellm/proxy/management_endpoints/mcp_management_endpoints.py b/litellm/proxy/management_endpoints/mcp_management_endpoints.py index 08f452859f..7eeb2a8300 100644 --- a/litellm/proxy/management_endpoints/mcp_management_endpoints.py +++ b/litellm/proxy/management_endpoints/mcp_management_endpoints.py @@ -412,6 +412,17 @@ if MCP_AVAILABLE: inherited_credentials["client_secret"] = existing_server.client_secret if existing_server.scopes: inherited_credentials["scopes"] = existing_server.scopes + # AWS SigV4 fields + if existing_server.aws_access_key_id: + inherited_credentials["aws_access_key_id"] = existing_server.aws_access_key_id + if existing_server.aws_secret_access_key: + inherited_credentials["aws_secret_access_key"] = existing_server.aws_secret_access_key + if existing_server.aws_session_token: + inherited_credentials["aws_session_token"] = existing_server.aws_session_token + if existing_server.aws_region_name: + inherited_credentials["aws_region_name"] = existing_server.aws_region_name + if existing_server.aws_service_name: + inherited_credentials["aws_service_name"] = existing_server.aws_service_name if not inherited_credentials: return payload diff --git a/litellm/proxy/pass_through_endpoints/pass_through_endpoints.py b/litellm/proxy/pass_through_endpoints/pass_through_endpoints.py index 356807415d..a06a3aa3da 100644 --- a/litellm/proxy/pass_through_endpoints/pass_through_endpoints.py +++ b/litellm/proxy/pass_through_endpoints/pass_through_endpoints.py @@ -404,7 +404,9 @@ class HttpPassThroughEndpointHelpers(BasePassthroughUtils): headers=headers, params=requested_query_params, ) - elif HttpPassThroughEndpointHelpers.is_multipart(request) is True: + elif HttpPassThroughEndpointHelpers.is_multipart(request) is True and not _parsed_body: + # Only use multipart handler if we don't have a parsed body + # (parsed body means it was JSON despite multipart content-type header) return await HttpPassThroughEndpointHelpers.make_multipart_http_request( request=request, async_client=async_client, @@ -677,8 +679,15 @@ async def pass_through_request( # noqa: PLR0915 str(url) ) + # Skip body parsing for multipart requests - make_multipart_http_request will handle it + # But if custom_body is provided (e.g., JSON parsed despite multipart content-type), use it + is_multipart = HttpPassThroughEndpointHelpers.is_multipart(request) and not custom_body + if custom_body: _parsed_body = custom_body + elif is_multipart: + # Don't parse multipart body here - it will be handled by make_multipart_http_request + _parsed_body = {} else: _parsed_body = await _read_request_body(request) verbose_proxy_logger.debug( @@ -1043,30 +1052,22 @@ async def _parse_request_data_by_content_type( # Handle requests with no body (e.g., DELETE requests) pass elif "multipart/form-data" in content_type: - # ✅ Handle multipart form-data - form = await request.form() - if "query_params" in form: - form_value = form["query_params"] - if isinstance(form_value, str): - try: - query_params_data = json.loads(form_value) - except Exception: - query_params_data = form_value - else: - query_params_data = form_value - - if "custom_body" in form: - form_value = form["custom_body"] - if isinstance(form_value, str): - try: - custom_body_data = json.loads(form_value) - except Exception: - custom_body_data = form_value - else: - custom_body_data = form_value - - if "file" in form: - file_data = form["file"] # this is a Starlette UploadFile object + # ✅ Try to parse as JSON first (handles misconfigured clients sending JSON with multipart content-type) + # If that fails, skip parsing - pass_through_request will handle actual multipart + try: + body = await request.json() + # Successfully parsed as JSON - treat as JSON body + query_params_data = body.get("query_params") + custom_body_data = body.get("custom_body") + stream = body.get("stream") + # If custom_body is not set, use the entire body + if custom_body_data is None and body: + custom_body_data = body + except (json.JSONDecodeError, Exception): + # Not JSON - this is actual multipart data + # Skip parsing here to avoid consuming the request body stream + # make_multipart_http_request will handle it + pass elif "application/x-www-form-urlencoded" in content_type: # ✅ Handle URL-encoded form data @@ -1132,7 +1133,6 @@ def create_pass_through_route( fastapi_response: Response, user_api_key_dict: UserAPIKeyAuth = Depends(user_api_key_auth), subpath: str = "", # captures sub-paths when include_subpath=True - custom_body: Optional[dict] = None, ): from litellm.proxy.pass_through_endpoints.pass_through_endpoints import ( InitPassThroughEndpointHelpers, @@ -1208,12 +1208,9 @@ def create_pass_through_route( ) if query_params: final_query_params.update(query_params) - # When a caller (e.g. bedrock_proxy_route) supplies a pre-built - # body, use it instead of the body parsed from the raw request. + # Use the body parsed from the raw request final_custom_body: Optional[dict] = None - if custom_body is not None: - final_custom_body = custom_body - elif isinstance(custom_body_data, dict): + if isinstance(custom_body_data, dict): final_custom_body = custom_body_data return await pass_through_request( # type: ignore diff --git a/litellm/proxy/proxy_cli.py b/litellm/proxy/proxy_cli.py index 921d86c35c..be2f5ac7c1 100644 --- a/litellm/proxy/proxy_cli.py +++ b/litellm/proxy/proxy_cli.py @@ -853,7 +853,12 @@ def run_server( # noqa: PLR0915 ): check_prisma_schema_diff(db_url=None) else: - PrismaManager.setup_database(use_migrate=not use_prisma_db_push) + if not PrismaManager.setup_database(use_migrate=not use_prisma_db_push): + print( # noqa + "\033[1;31mLiteLLM Proxy: Database setup failed after multiple retries. " + "The proxy cannot start safely. Please check your database connection and migration status.\033[0m" + ) + sys.exit(1) else: print( # noqa f"Unable to connect to DB. DATABASE_URL found in environment, but prisma package not found." # noqa diff --git a/litellm/types/llms/vertex_ai.py b/litellm/types/llms/vertex_ai.py index 45af926b0b..d94ba5f805 100644 --- a/litellm/types/llms/vertex_ai.py +++ b/litellm/types/llms/vertex_ai.py @@ -557,6 +557,17 @@ class VertexAIBatchEmbeddingsResponseObject(TypedDict): embeddings: List[ContentEmbeddings] +class GeminiEmbedContentRequestBody(TypedDict, total=False): + content: Required[ContentType] + taskType: TaskTypeEnum + title: str + outputDimensionality: int + + +class GeminiEmbedContentResponseObject(TypedDict): + embedding: ContentEmbeddings + + # Vertex AI Batch Prediction diff --git a/litellm/types/mcp.py b/litellm/types/mcp.py index 33e55f9bed..af91926de2 100644 --- a/litellm/types/mcp.py +++ b/litellm/types/mcp.py @@ -95,6 +95,22 @@ class MCPCredentials(TypedDict, total=False): OAuth 2.0 scopes to request when exchanging the client credentials """ + # AWS SigV4 fields + aws_access_key_id: Optional[str] + """AWS access key ID for SigV4 signing. Optional — falls back to boto3 credential chain.""" + + aws_secret_access_key: Optional[str] + """AWS secret access key for SigV4 signing. Optional — falls back to boto3 credential chain.""" + + aws_session_token: Optional[str] + """AWS session token for temporary STS credentials. Optional.""" + + aws_region_name: Optional[str] + """AWS region for SigV4 signing (e.g., 'us-east-1'). Not a secret — stored unencrypted.""" + + aws_service_name: Optional[str] + """AWS service name for SigV4 signing (e.g., 'bedrock-agentcore'). Not a secret — stored unencrypted.""" + class MCPServerCostInfo(TypedDict, total=False): default_cost_per_query: Optional[float] diff --git a/litellm/types/utils.py b/litellm/types/utils.py index c163ef6b07..341e5117fd 100644 --- a/litellm/types/utils.py +++ b/litellm/types/utils.py @@ -253,6 +253,7 @@ class ModelInfoBase(ProviderSpecificModelInfo, total=False): tpm: Optional[int] rpm: Optional[int] provider_specific_entry: Optional[Dict[str, float]] + uses_embed_content: Optional[bool] class ModelInfo(ModelInfoBase, total=False): diff --git a/litellm/utils.py b/litellm/utils.py index a86447be03..88312354bc 100644 --- a/litellm/utils.py +++ b/litellm/utils.py @@ -5779,6 +5779,7 @@ def _get_model_info_helper( # noqa: PLR0915 provider_specific_entry=_model_info.get( "provider_specific_entry", None ), + uses_embed_content=_model_info.get("uses_embed_content", None), ) except Exception as e: verbose_logger.debug(f"Error getting model info: {e}") diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json index 5b171f65d5..3006d91f78 100644 --- a/model_prices_and_context_window.json +++ b/model_prices_and_context_window.json @@ -1,4 +1,44 @@ { + "sample_spec": { + "code_interpreter_cost_per_session": 0.0, + "computer_use_input_cost_per_1k_tokens": 0.0, + "computer_use_output_cost_per_1k_tokens": 0.0, + "deprecation_date": "date when the model becomes deprecated in the format YYYY-MM-DD", + "file_search_cost_per_1k_calls": 0.0, + "file_search_cost_per_gb_per_day": 0.0, + "input_cost_per_audio_token": 0.0, + "input_cost_per_token": 0.0, + "litellm_provider": "one of https://docs.litellm.ai/docs/providers", + "max_input_tokens": "max input tokens, if the provider specifies it. if not default to max_tokens", + "max_output_tokens": "max output tokens, if the provider specifies it. if not default to max_tokens", + "max_tokens": "LEGACY parameter. set to max_output_tokens if provider specifies it. IF not set to max_input_tokens, if provider specifies it.", + "mode": "one of: chat, embedding, completion, image_generation, audio_transcription, audio_speech, image_generation, moderation, rerank, search", + "output_cost_per_reasoning_token": 0.0, + "output_cost_per_token": 0.0, + "search_context_cost_per_query": { + "search_context_size_high": 0.0, + "search_context_size_low": 0.0, + "search_context_size_medium": 0.0 + }, + "supported_regions": [ + "global", + "us-west-2", + "eu-west-1", + "ap-southeast-1", + "ap-northeast-1" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_vision": true, + "supports_web_search": true, + "vector_store_cost_per_gb_per_day": 0.0 + }, "1024-x-1024/50-steps/bedrock/amazon.nova-canvas-v1:0": { "litellm_provider": "bedrock", "max_input_tokens": 2600, @@ -18,28 +58,6 @@ "mode": "image_generation", "output_cost_per_pixel": 0.0 }, - "1024-x-1024/gpt-image-1.5": { - "input_cost_per_image": 0.009, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "1024-x-1024/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.009, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, "1024-x-1024/max-steps/stability.stable-diffusion-xl-v1": { "litellm_provider": "bedrock", "max_input_tokens": 77, @@ -47,50 +65,6 @@ "mode": "image_generation", "output_cost_per_image": 0.08 }, - "1024-x-1536/gpt-image-1.5": { - "input_cost_per_image": 0.013, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "1024-x-1536/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.013, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "1536-x-1024/gpt-image-1.5": { - "input_cost_per_image": 0.013, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "1536-x-1024/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.013, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, "256-x-256/dall-e-2": { "input_cost_per_pixel": 2.4414e-07, "litellm_provider": "openai", @@ -299,9 +273,59 @@ "/v1/images/generations" ] }, - "amazon-nova/nova-lite-v1": { + "amazon.nova-canvas-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 2600, + "mode": "image_generation", + "output_cost_per_image": 0.06 + }, + "us.writer.palmyra-x4-v1:0": { + "input_cost_per_token": 2.5e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 128000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 1e-05, + "supports_function_calling": true, + "supports_pdf_input": true + }, + "us.writer.palmyra-x5-v1:0": { + "input_cost_per_token": 6e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 6e-06, + "supports_function_calling": true, + "supports_pdf_input": true + }, + "writer.palmyra-x4-v1:0": { + "input_cost_per_token": 2.5e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 128000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 1e-05, + "supports_function_calling": true, + "supports_pdf_input": true + }, + "writer.palmyra-x5-v1:0": { + "input_cost_per_token": 6e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 6e-06, + "supports_function_calling": true, + "supports_pdf_input": true + }, + "amazon.nova-lite-v1:0": { "input_cost_per_token": 6e-08, - "litellm_provider": "amazon_nova", + "litellm_provider": "bedrock_converse", "max_input_tokens": 300000, "max_output_tokens": 10000, "max_tokens": 10000, @@ -313,46 +337,6 @@ "supports_response_schema": true, "supports_vision": true }, - "amazon-nova/nova-micro-v1": { - "input_cost_per_token": 3.5e-08, - "litellm_provider": "amazon_nova", - "max_input_tokens": 128000, - "max_output_tokens": 10000, - "max_tokens": 10000, - "mode": "chat", - "output_cost_per_token": 1.4e-07, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_response_schema": true - }, - "amazon-nova/nova-premier-v1": { - "input_cost_per_token": 2.5e-06, - "litellm_provider": "amazon_nova", - "max_input_tokens": 1000000, - "max_output_tokens": 10000, - "max_tokens": 10000, - "mode": "chat", - "output_cost_per_token": 1.25e-05, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": false, - "supports_response_schema": true, - "supports_vision": true - }, - "amazon-nova/nova-pro-v1": { - "input_cost_per_token": 8e-07, - "litellm_provider": "amazon_nova", - "max_input_tokens": 300000, - "max_output_tokens": 10000, - "max_tokens": 10000, - "mode": "chat", - "output_cost_per_token": 3.2e-06, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_vision": true - }, "amazon.nova-2-lite-v1:0": { "cache_read_input_token_cost": 7.5e-08, "input_cost_per_token": 3e-07, @@ -370,6 +354,133 @@ "supports_video_input": true, "supports_vision": true }, + "amazon.nova-2-pro-preview-20251202-v1:0": { + "cache_read_input_token_cost": 5.46875e-07, + "input_cost_per_token": 2.1875e-06, + "input_cost_per_image_token": 2.1875e-06, + "input_cost_per_audio_token": 2.1875e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.75e-05, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_video_input": true, + "supports_vision": true + }, + "apac.amazon.nova-2-lite-v1:0": { + "cache_read_input_token_cost": 8.25e-08, + "input_cost_per_token": 3.3e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 2.75e-06, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_video_input": true, + "supports_vision": true + }, + "apac.amazon.nova-2-pro-preview-20251202-v1:0": { + "cache_read_input_token_cost": 5.46875e-07, + "input_cost_per_token": 2.1875e-06, + "input_cost_per_image_token": 2.1875e-06, + "input_cost_per_audio_token": 2.1875e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.75e-05, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_video_input": true, + "supports_vision": true + }, + "eu.amazon.nova-2-lite-v1:0": { + "cache_read_input_token_cost": 8.25e-08, + "input_cost_per_token": 3.3e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 2.75e-06, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_video_input": true, + "supports_vision": true + }, + "eu.amazon.nova-2-pro-preview-20251202-v1:0": { + "cache_read_input_token_cost": 5.46875e-07, + "input_cost_per_token": 2.1875e-06, + "input_cost_per_image_token": 2.1875e-06, + "input_cost_per_audio_token": 2.1875e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.75e-05, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_video_input": true, + "supports_vision": true + }, + "us.amazon.nova-2-lite-v1:0": { + "cache_read_input_token_cost": 8.25e-08, + "input_cost_per_token": 3.3e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 2.75e-06, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_video_input": true, + "supports_vision": true + }, + "us.amazon.nova-2-pro-preview-20251202-v1:0": { + "cache_read_input_token_cost": 5.46875e-07, + "input_cost_per_token": 2.1875e-06, + "input_cost_per_image_token": 2.1875e-06, + "input_cost_per_audio_token": 2.1875e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.75e-05, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_video_input": true, + "supports_vision": true + }, "amazon.nova-2-multimodal-embeddings-v1:0": { "litellm_provider": "bedrock", "max_input_tokens": 8172, @@ -387,45 +498,6 @@ "supports_video_input": true, "supports_audio_input": true }, - "amazon.nova-2-pro-preview-20251202-v1:0": { - "cache_read_input_token_cost": 5.46875e-07, - "input_cost_per_token": 2.1875e-06, - "input_cost_per_image_token": 2.1875e-06, - "input_cost_per_audio_token": 2.1875e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.75e-05, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_video_input": true, - "supports_vision": true - }, - "amazon.nova-canvas-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 2600, - "mode": "image_generation", - "output_cost_per_image": 0.06 - }, - "amazon.nova-lite-v1:0": { - "input_cost_per_token": 6e-08, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 300000, - "max_output_tokens": 10000, - "max_tokens": 10000, - "mode": "chat", - "output_cost_per_token": 2.4e-07, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_vision": true - }, "amazon.nova-micro-v1:0": { "input_cost_per_token": 3.5e-08, "litellm_provider": "bedrock_converse", @@ -526,6 +598,66 @@ "litellm_provider": "bedrock", "mode": "image_generation" }, + "twelvelabs.marengo-embed-2-7-v1:0": { + "input_cost_per_token": 7e-05, + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "max_tokens": 77, + "mode": "embedding", + "output_cost_per_token": 0.0, + "output_vector_size": 1024, + "supports_embedding_image_input": true, + "supports_image_input": true + }, + "us.twelvelabs.marengo-embed-2-7-v1:0": { + "input_cost_per_token": 7e-05, + "input_cost_per_video_per_second": 0.0007, + "input_cost_per_audio_per_second": 0.00014, + "input_cost_per_image": 0.0001, + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "max_tokens": 77, + "mode": "embedding", + "output_cost_per_token": 0.0, + "output_vector_size": 1024, + "supports_embedding_image_input": true, + "supports_image_input": true + }, + "eu.twelvelabs.marengo-embed-2-7-v1:0": { + "input_cost_per_token": 7e-05, + "input_cost_per_video_per_second": 0.0007, + "input_cost_per_audio_per_second": 0.00014, + "input_cost_per_image": 0.0001, + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "max_tokens": 77, + "mode": "embedding", + "output_cost_per_token": 0.0, + "output_vector_size": 1024, + "supports_embedding_image_input": true, + "supports_image_input": true + }, + "twelvelabs.pegasus-1-2-v1:0": { + "input_cost_per_video_per_second": 0.00049, + "output_cost_per_token": 7.5e-06, + "litellm_provider": "bedrock", + "mode": "chat", + "supports_video_input": true + }, + "us.twelvelabs.pegasus-1-2-v1:0": { + "input_cost_per_video_per_second": 0.00049, + "output_cost_per_token": 7.5e-06, + "litellm_provider": "bedrock", + "mode": "chat", + "supports_video_input": true + }, + "eu.twelvelabs.pegasus-1-2-v1:0": { + "input_cost_per_video_per_second": 0.00049, + "output_cost_per_token": 7.5e-06, + "litellm_provider": "bedrock", + "mode": "chat", + "supports_video_input": true + }, "amazon.titan-text-express-v1": { "input_cost_per_token": 1.3e-06, "litellm_provider": "bedrock", @@ -570,6 +702,51 @@ "supports_response_schema": true, "supports_tool_choice": true }, + "anthropic.claude-haiku-4-5-20251001-v1:0": { + "cache_creation_input_token_cost": 1.25e-06, + "cache_read_input_token_cost": 1e-07, + "input_cost_per_token": 1e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 5e-06, + "source": "https://aws.amazon.com/about-aws/whats-new/2025/10/claude-4-5-haiku-anthropic-amazon-bedrock", + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "anthropic.claude-haiku-4-5@20251001": { + "cache_creation_input_token_cost": 1.25e-06, + "cache_read_input_token_cost": 1e-07, + "input_cost_per_token": 1e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 5e-06, + "source": "https://aws.amazon.com/about-aws/whats-new/2025/10/claude-4-5-haiku-anthropic-amazon-bedrock", + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346, + "supports_native_streaming": true + }, "anthropic.claude-3-5-sonnet-20240620-v1:0": { "input_cost_per_token": 3e-06, "litellm_provider": "bedrock", @@ -704,51 +881,6 @@ "cache_read_input_token_cost": 3e-07, "cache_creation_input_token_cost": 3.75e-06 }, - "anthropic.claude-haiku-4-5-20251001-v1:0": { - "cache_creation_input_token_cost": 1.25e-06, - "cache_read_input_token_cost": 1e-07, - "input_cost_per_token": 1e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 5e-06, - "source": "https://aws.amazon.com/about-aws/whats-new/2025/10/claude-4-5-haiku-anthropic-amazon-bedrock", - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, - "anthropic.claude-haiku-4-5@20251001": { - "cache_creation_input_token_cost": 1.25e-06, - "cache_read_input_token_cost": 1e-07, - "input_cost_per_token": 1e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 5e-06, - "source": "https://aws.amazon.com/about-aws/whats-new/2025/10/claude-4-5-haiku-anthropic-amazon-bedrock", - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346, - "supports_native_streaming": true - }, "anthropic.claude-instant-v1": { "input_cost_per_token": 8e-07, "litellm_provider": "bedrock", @@ -867,6 +999,276 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 346 }, + "global.anthropic.claude-opus-4-6-v1": { + "cache_creation_input_token_cost": 6.25e-06, + "cache_creation_input_token_cost_above_200k_tokens": 1.25e-05, + "cache_read_input_token_cost": 5e-07, + "cache_read_input_token_cost_above_200k_tokens": 1e-06, + "input_cost_per_token": 5e-06, + "input_cost_per_token_above_200k_tokens": 1e-05, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 2.5e-05, + "output_cost_per_token_above_200k_tokens": 3.75e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": false, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "us.anthropic.claude-opus-4-6-v1": { + "cache_creation_input_token_cost": 6.875e-06, + "cache_creation_input_token_cost_above_200k_tokens": 1.375e-05, + "cache_read_input_token_cost": 5.5e-07, + "cache_read_input_token_cost_above_200k_tokens": 1.1e-06, + "input_cost_per_token": 5.5e-06, + "input_cost_per_token_above_200k_tokens": 1.1e-05, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 2.75e-05, + "output_cost_per_token_above_200k_tokens": 4.125e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": false, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "eu.anthropic.claude-opus-4-6-v1": { + "cache_creation_input_token_cost": 6.875e-06, + "cache_creation_input_token_cost_above_200k_tokens": 1.375e-05, + "cache_read_input_token_cost": 5.5e-07, + "cache_read_input_token_cost_above_200k_tokens": 1.1e-06, + "input_cost_per_token": 5.5e-06, + "input_cost_per_token_above_200k_tokens": 1.1e-05, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 2.75e-05, + "output_cost_per_token_above_200k_tokens": 4.125e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": false, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "au.anthropic.claude-opus-4-6-v1": { + "cache_creation_input_token_cost": 6.875e-06, + "cache_creation_input_token_cost_above_200k_tokens": 1.375e-05, + "cache_read_input_token_cost": 5.5e-07, + "cache_read_input_token_cost_above_200k_tokens": 1.1e-06, + "input_cost_per_token": 5.5e-06, + "input_cost_per_token_above_200k_tokens": 1.1e-05, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 2.75e-05, + "output_cost_per_token_above_200k_tokens": 4.125e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": false, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "anthropic.claude-sonnet-4-6": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost": 3e-07, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "global.anthropic.claude-sonnet-4-6": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost": 3e-07, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "us.anthropic.claude-sonnet-4-6": { + "cache_creation_input_token_cost": 4.125e-06, + "cache_creation_input_token_cost_above_200k_tokens": 8.25e-06, + "cache_read_input_token_cost": 3.3e-07, + "cache_read_input_token_cost_above_200k_tokens": 6.6e-07, + "input_cost_per_token": 3.3e-06, + "input_cost_per_token_above_200k_tokens": 6.6e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.65e-05, + "output_cost_per_token_above_200k_tokens": 2.475e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "eu.anthropic.claude-sonnet-4-6": { + "cache_creation_input_token_cost": 4.125e-06, + "cache_creation_input_token_cost_above_200k_tokens": 8.25e-06, + "cache_read_input_token_cost": 3.3e-07, + "cache_read_input_token_cost_above_200k_tokens": 6.6e-07, + "input_cost_per_token": 3.3e-06, + "input_cost_per_token_above_200k_tokens": 6.6e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.65e-05, + "output_cost_per_token_above_200k_tokens": 2.475e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "au.anthropic.claude-sonnet-4-6": { + "cache_creation_input_token_cost": 4.125e-06, + "cache_creation_input_token_cost_above_200k_tokens": 8.25e-06, + "cache_read_input_token_cost": 3.3e-07, + "cache_read_input_token_cost_above_200k_tokens": 6.6e-07, + "input_cost_per_token": 3.3e-06, + "input_cost_per_token_above_200k_tokens": 6.6e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.65e-05, + "output_cost_per_token_above_200k_tokens": 2.475e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, "anthropic.claude-sonnet-4-20250514-v1:0": { "cache_creation_input_token_cost": 3.75e-06, "cache_read_input_token_cost": 3e-07, @@ -927,36 +1329,6 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, - "anthropic.claude-sonnet-4-6": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost": 3e-07, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, "anthropic.claude-v1": { "input_cost_per_token": 8e-06, "litellm_provider": "bedrock", @@ -1094,42 +1466,6 @@ "source": "https://docs.anyscale.com/preview/endpoints/text-generation/supported-models/mistralai-Mixtral-8x7B-Instruct-v0.1", "supports_function_calling": true }, - "apac.amazon.nova-2-lite-v1:0": { - "cache_read_input_token_cost": 8.25e-08, - "input_cost_per_token": 3.3e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 2.75e-06, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_video_input": true, - "supports_vision": true - }, - "apac.amazon.nova-2-pro-preview-20251202-v1:0": { - "cache_read_input_token_cost": 5.46875e-07, - "input_cost_per_token": 2.1875e-06, - "input_cost_per_image_token": 2.1875e-06, - "input_cost_per_audio_token": 2.1875e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.75e-05, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_video_input": true, - "supports_vision": true - }, "apac.amazon.nova-lite-v1:0": { "input_cost_per_token": 6.3e-08, "litellm_provider": "bedrock_converse", @@ -1221,22 +1557,6 @@ "cache_read_input_token_cost": 2.5e-08, "cache_creation_input_token_cost": 3.125e-07 }, - "apac.anthropic.claude-3-sonnet-20240229-v1:0": { - "input_cost_per_token": 3e-06, - "litellm_provider": "bedrock", - "max_input_tokens": 200000, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "cache_read_input_token_cost": 3e-07, - "cache_creation_input_token_cost": 3.75e-06 - }, "apac.anthropic.claude-haiku-4-5-20251001-v1:0": { "cache_creation_input_token_cost": 1.375e-06, "cache_read_input_token_cost": 1.1e-07, @@ -1259,6 +1579,22 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 346 }, + "apac.anthropic.claude-3-sonnet-20240229-v1:0": { + "input_cost_per_token": 3e-06, + "litellm_provider": "bedrock", + "max_input_tokens": 200000, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "cache_read_input_token_cost": 3e-07, + "cache_creation_input_token_cost": 3.75e-06 + }, "apac.anthropic.claude-sonnet-4-20250514-v1:0": { "cache_creation_input_token_cost": 3.75e-06, "cache_read_input_token_cost": 3e-07, @@ -1301,57 +1637,6 @@ "mode": "audio_transcription", "output_cost_per_second": 0.0 }, - "au.anthropic.claude-haiku-4-5-20251001-v1:0": { - "cache_creation_input_token_cost": 1.375e-06, - "cache_read_input_token_cost": 1.1e-07, - "input_cost_per_token": 1.1e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 5.5e-06, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, - "au.anthropic.claude-opus-4-6-v1": { - "cache_creation_input_token_cost": 6.875e-06, - "cache_creation_input_token_cost_above_200k_tokens": 1.375e-05, - "cache_read_input_token_cost": 5.5e-07, - "cache_read_input_token_cost_above_200k_tokens": 1.1e-06, - "input_cost_per_token": 5.5e-06, - "input_cost_per_token_above_200k_tokens": 1.1e-05, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 2.75e-05, - "output_cost_per_token_above_200k_tokens": 4.125e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": false, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, "au.anthropic.claude-sonnet-4-5-20250929-v1:0": { "cache_creation_input_token_cost": 4.125e-06, "cache_read_input_token_cost": 3.3e-07, @@ -1382,72 +1667,6 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 346 }, - "au.anthropic.claude-sonnet-4-6": { - "cache_creation_input_token_cost": 4.125e-06, - "cache_creation_input_token_cost_above_200k_tokens": 8.25e-06, - "cache_read_input_token_cost": 3.3e-07, - "cache_read_input_token_cost_above_200k_tokens": 6.6e-07, - "input_cost_per_token": 3.3e-06, - "input_cost_per_token_above_200k_tokens": 6.6e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.65e-05, - "output_cost_per_token_above_200k_tokens": 2.475e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, - "aws_polly/generative": { - "input_cost_per_character": 3e-05, - "litellm_provider": "aws_polly", - "mode": "audio_speech", - "supported_endpoints": [ - "/v1/audio/speech" - ], - "source": "https://aws.amazon.com/polly/pricing/" - }, - "aws_polly/long-form": { - "input_cost_per_character": 0.0001, - "litellm_provider": "aws_polly", - "mode": "audio_speech", - "supported_endpoints": [ - "/v1/audio/speech" - ], - "source": "https://aws.amazon.com/polly/pricing/" - }, - "aws_polly/neural": { - "input_cost_per_character": 1.6e-05, - "litellm_provider": "aws_polly", - "mode": "audio_speech", - "supported_endpoints": [ - "/v1/audio/speech" - ], - "source": "https://aws.amazon.com/polly/pricing/" - }, - "aws_polly/standard": { - "input_cost_per_character": 4e-06, - "litellm_provider": "aws_polly", - "mode": "audio_speech", - "supported_endpoints": [ - "/v1/audio/speech" - ], - "source": "https://aws.amazon.com/polly/pricing/" - }, "azure/ada": { "input_cost_per_token": 1e-07, "litellm_provider": "azure", @@ -1495,6 +1714,139 @@ "output_cost_per_token": 1.5e-05, "supports_function_calling": true }, + "azure_ai/claude-haiku-4-5": { + "cache_creation_input_token_cost": 1.25e-06, + "cache_creation_input_token_cost_above_1hr": 2e-06, + "cache_read_input_token_cost": 1e-07, + "input_cost_per_token": 1e-06, + "litellm_provider": "azure_ai", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 5e-06, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "azure_ai/claude-opus-4-5": { + "cache_creation_input_token_cost": 6.25e-06, + "cache_creation_input_token_cost_above_1hr": 1e-05, + "cache_read_input_token_cost": 5e-07, + "input_cost_per_token": 5e-06, + "litellm_provider": "azure_ai", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 2.5e-05, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "azure_ai/claude-opus-4-6": { + "input_cost_per_token": 5e-06, + "output_cost_per_token": 2.5e-05, + "litellm_provider": "azure_ai", + "max_input_tokens": 200000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "cache_creation_input_token_cost": 6.25e-06, + "cache_creation_input_token_cost_above_1hr": 1e-05, + "cache_read_input_token_cost": 5e-07, + "supports_assistant_prefill": false, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 159 + }, + "azure_ai/claude-opus-4-1": { + "cache_creation_input_token_cost": 1.875e-05, + "cache_creation_input_token_cost_above_1hr": 3e-05, + "cache_read_input_token_cost": 1.5e-06, + "input_cost_per_token": 1.5e-05, + "litellm_provider": "azure_ai", + "max_input_tokens": 200000, + "max_output_tokens": 32000, + "max_tokens": 32000, + "mode": "chat", + "output_cost_per_token": 7.5e-05, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "azure_ai/claude-sonnet-4-5": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_creation_input_token_cost_above_1hr": 6e-06, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 3e-06, + "litellm_provider": "azure_ai", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "azure_ai/claude-sonnet-4-6": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_creation_input_token_cost_above_1hr": 6e-06, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 3e-06, + "litellm_provider": "azure_ai", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, "azure/computer-use-preview": { "input_cost_per_token": 3e-06, "litellm_provider": "azure", @@ -1527,6 +1879,28 @@ "litellm_provider": "azure", "mode": "chat" }, + "azure_ai/gpt-oss-120b": { + "input_cost_per_token": 1.5e-07, + "output_cost_per_token": 6e-07, + "litellm_provider": "azure_ai", + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "max_tokens": 131072, + "mode": "chat", + "source": "https://azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "azure_ai/model_router": { + "input_cost_per_token": 1.4e-07, + "output_cost_per_token": 0, + "litellm_provider": "azure_ai", + "mode": "chat", + "source": "https://azure.microsoft.com/en-us/pricing/details/ai-services/", + "comment": "Flat cost of $0.14 per M input tokens for Azure AI Foundry Model Router infrastructure. Use pattern: azure_ai/model_router/ where deployment-name is your Azure deployment (e.g., azure-model-router)" + }, "azure/eu/gpt-4o-2024-08-06": { "deprecation_date": "2026-02-27", "cache_read_input_token_cost": 1.375e-06, @@ -1705,38 +2079,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "azure/eu/gpt-5-nano-2025-08-07": { - "cache_read_input_token_cost": 5.5e-09, - "input_cost_per_token": 5.5e-08, - "litellm_provider": "azure", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 4.4e-07, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true - }, "azure/eu/gpt-5.1": { "cache_read_input_token_cost": 1.4e-07, "input_cost_per_token": 1.38e-06, @@ -1865,6 +2207,38 @@ "supports_tool_choice": true, "supports_vision": true }, + "azure/eu/gpt-5-nano-2025-08-07": { + "cache_read_input_token_cost": 5.5e-09, + "input_cost_per_token": 5.5e-08, + "litellm_provider": "azure", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 4.4e-07, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true + }, "azure/eu/o1-2024-12-17": { "cache_read_input_token_cost": 8.25e-06, "input_cost_per_token": 1.65e-05, @@ -2625,6 +2999,99 @@ "supports_tool_choice": true, "supports_vision": true }, + "azure/gpt-audio-2025-08-28": { + "input_cost_per_audio_token": 4e-05, + "input_cost_per_token": 2.5e-06, + "litellm_provider": "azure", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_audio_token": 8e-05, + "output_cost_per_token": 1e-05, + "supported_endpoints": [ + "/v1/chat/completions" + ], + "supported_modalities": [ + "text", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_prompt_caching": false, + "supports_reasoning": false, + "supports_response_schema": false, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": false + }, + "azure/gpt-audio-1.5-2026-02-23": { + "input_cost_per_audio_token": 4e-05, + "input_cost_per_token": 2.5e-06, + "litellm_provider": "azure", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_audio_token": 8e-05, + "output_cost_per_token": 1e-05, + "supported_endpoints": [ + "/v1/chat/completions" + ], + "supported_modalities": [ + "text", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_prompt_caching": false, + "supports_reasoning": false, + "supports_response_schema": false, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": false + }, + "azure/gpt-audio-mini-2025-10-06": { + "input_cost_per_audio_token": 1e-05, + "input_cost_per_token": 6e-07, + "litellm_provider": "azure", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_audio_token": 2e-05, + "output_cost_per_token": 2.4e-06, + "supported_endpoints": [ + "/v1/chat/completions" + ], + "supported_modalities": [ + "text", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_prompt_caching": false, + "supports_reasoning": false, + "supports_response_schema": false, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": false + }, "azure/gpt-4o-audio-preview-2024-12-17": { "input_cost_per_audio_token": 4e-05, "input_cost_per_token": 2.5e-06, @@ -2738,6 +3205,102 @@ "supports_system_messages": true, "supports_tool_choice": true }, + "azure/gpt-realtime-2025-08-28": { + "cache_creation_input_audio_token_cost": 4e-06, + "cache_read_input_token_cost": 4e-06, + "input_cost_per_audio_token": 3.2e-05, + "input_cost_per_image": 5e-06, + "input_cost_per_token": 4e-06, + "litellm_provider": "azure", + "max_input_tokens": 32000, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_audio_token": 6.4e-05, + "output_cost_per_token": 1.6e-05, + "supported_endpoints": [ + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "image", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "azure/gpt-realtime-1.5-2026-02-23": { + "cache_creation_input_audio_token_cost": 4e-06, + "cache_read_input_token_cost": 4e-06, + "input_cost_per_audio_token": 3.2e-05, + "input_cost_per_image": 5e-06, + "input_cost_per_token": 4e-06, + "litellm_provider": "azure", + "max_input_tokens": 32000, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_audio_token": 6.4e-05, + "output_cost_per_token": 1.6e-05, + "supported_endpoints": [ + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "image", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "azure/gpt-realtime-mini-2025-10-06": { + "cache_creation_input_audio_token_cost": 3e-07, + "cache_read_input_token_cost": 6e-08, + "input_cost_per_audio_token": 1e-05, + "input_cost_per_image": 8e-07, + "input_cost_per_token": 6e-07, + "litellm_provider": "azure", + "max_input_tokens": 32000, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_audio_token": 2e-05, + "output_cost_per_token": 2.4e-06, + "supported_endpoints": [ + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "image", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, "azure/gpt-4o-mini-transcribe": { "input_cost_per_audio_token": 3e-06, "input_cost_per_token": 1.25e-06, @@ -2837,6 +3400,145 @@ "/v1/audio/transcriptions" ] }, + "azure/gpt-5.1-2025-11-13": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_priority": 2.5e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_priority": 2.5e-06, + "litellm_provider": "azure", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_priority": 2e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true, + "supports_none_reasoning_effort": true + }, + "azure/gpt-5.1-chat-2025-11-13": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_priority": 2.5e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_priority": 2.5e-06, + "litellm_provider": "azure", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_priority": 2e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": false, + "supports_native_streaming": true, + "supports_parallel_function_calling": false, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": false, + "supports_vision": true, + "supports_none_reasoning_effort": true + }, + "azure/gpt-5.1-codex-2025-11-13": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_priority": 2.5e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_priority": 2.5e-06, + "litellm_provider": "azure", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 1e-05, + "output_cost_per_token_priority": 2e-05, + "supported_endpoints": [ + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": false, + "supports_tool_choice": true, + "supports_vision": true + }, + "azure/gpt-5.1-codex-mini-2025-11-13": { + "cache_read_input_token_cost": 2.5e-08, + "cache_read_input_token_cost_priority": 4.5e-08, + "input_cost_per_token": 2.5e-07, + "input_cost_per_token_priority": 4.5e-07, + "litellm_provider": "azure", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 2e-06, + "output_cost_per_token_priority": 3.6e-06, + "supported_endpoints": [ + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": false, + "supports_tool_choice": true, + "supports_vision": true + }, "azure/gpt-5": { "cache_read_input_token_cost": 1.25e-07, "input_cost_per_token": 1.25e-06, @@ -3187,43 +3889,6 @@ "supports_vision": true, "supports_none_reasoning_effort": true }, - "azure/gpt-5.1-2025-11-13": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_priority": 2.5e-07, - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_priority": 2.5e-06, - "litellm_provider": "azure", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_priority": 2e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true, - "supports_none_reasoning_effort": true - }, "azure/gpt-5.1-chat": { "cache_read_input_token_cost": 1.25e-07, "input_cost_per_token": 1.25e-06, @@ -3258,42 +3923,6 @@ "supports_vision": true, "supports_none_reasoning_effort": true }, - "azure/gpt-5.1-chat-2025-11-13": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_priority": 2.5e-07, - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_priority": 2.5e-06, - "litellm_provider": "azure", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_priority": 2e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_function_calling": false, - "supports_native_streaming": true, - "supports_parallel_function_calling": false, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": false, - "supports_vision": true, - "supports_none_reasoning_effort": true - }, "azure/gpt-5.1-codex": { "cache_read_input_token_cost": 1.25e-07, "input_cost_per_token": 1.25e-06, @@ -3324,39 +3953,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "azure/gpt-5.1-codex-2025-11-13": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_priority": 2.5e-07, - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_priority": 2.5e-06, - "litellm_provider": "azure", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 1e-05, - "output_cost_per_token_priority": 2e-05, - "supported_endpoints": [ - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": false, - "supports_tool_choice": true, - "supports_vision": true - }, "azure/gpt-5.1-codex-max": { "cache_read_input_token_cost": 1.25e-07, "input_cost_per_token": 1.25e-06, @@ -3417,39 +4013,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "azure/gpt-5.1-codex-mini-2025-11-13": { - "cache_read_input_token_cost": 2.5e-08, - "cache_read_input_token_cost_priority": 4.5e-08, - "input_cost_per_token": 2.5e-07, - "input_cost_per_token_priority": 4.5e-07, - "litellm_provider": "azure", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 2e-06, - "output_cost_per_token_priority": 3.6e-06, - "supported_endpoints": [ - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": false, - "supports_tool_choice": true, - "supports_vision": true - }, "azure/gpt-5.2": { "cache_read_input_token_cost": 1.75e-07, "input_cost_per_token": 1.75e-06, @@ -3618,68 +4181,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "azure/gpt-5.2-pro": { - "input_cost_per_token": 2.1e-05, - "litellm_provider": "azure", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 0.000168, - "supported_endpoints": [ - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true - }, - "azure/gpt-5.2-pro-2025-12-11": { - "input_cost_per_token": 2.1e-05, - "litellm_provider": "azure", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 0.000168, - "supported_endpoints": [ - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true - }, "azure/gpt-5.3-chat": { "cache_read_input_token_cost": 1.75e-07, "cache_read_input_token_cost_priority": 3.5e-07, @@ -3745,6 +4246,68 @@ "supports_tool_choice": true, "supports_vision": true }, + "azure/gpt-5.2-pro": { + "input_cost_per_token": 2.1e-05, + "litellm_provider": "azure", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 0.000168, + "supported_endpoints": [ + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true + }, + "azure/gpt-5.2-pro-2025-12-11": { + "input_cost_per_token": 2.1e-05, + "litellm_provider": "azure", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 0.000168, + "supported_endpoints": [ + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true + }, "azure/gpt-5.4": { "cache_read_input_token_cost": 2.5e-07, "cache_read_input_token_cost_above_272k_tokens": 5e-07, @@ -3899,99 +4462,6 @@ "supports_vision": true, "supports_web_search": true }, - "azure/gpt-audio-1.5-2026-02-23": { - "input_cost_per_audio_token": 4e-05, - "input_cost_per_token": 2.5e-06, - "litellm_provider": "azure", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_audio_token": 8e-05, - "output_cost_per_token": 1e-05, - "supported_endpoints": [ - "/v1/chat/completions" - ], - "supported_modalities": [ - "text", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_prompt_caching": false, - "supports_reasoning": false, - "supports_response_schema": false, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": false - }, - "azure/gpt-audio-2025-08-28": { - "input_cost_per_audio_token": 4e-05, - "input_cost_per_token": 2.5e-06, - "litellm_provider": "azure", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_audio_token": 8e-05, - "output_cost_per_token": 1e-05, - "supported_endpoints": [ - "/v1/chat/completions" - ], - "supported_modalities": [ - "text", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_prompt_caching": false, - "supports_reasoning": false, - "supports_response_schema": false, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": false - }, - "azure/gpt-audio-mini-2025-10-06": { - "input_cost_per_audio_token": 1e-05, - "input_cost_per_token": 6e-07, - "litellm_provider": "azure", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_audio_token": 2e-05, - "output_cost_per_token": 2.4e-06, - "supported_endpoints": [ - "/v1/chat/completions" - ], - "supported_modalities": [ - "text", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_prompt_caching": false, - "supports_reasoning": false, - "supports_response_schema": false, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": false - }, "azure/gpt-image-1": { "cache_read_input_image_token_cost": 2.5e-06, "cache_read_input_token_cost": 1.25e-06, @@ -4005,6 +4475,105 @@ "/v1/images/edits" ] }, + "azure/hd/1024-x-1024/dall-e-3": { + "input_cost_per_pixel": 7.629e-08, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_token": 0.0 + }, + "azure/hd/1024-x-1792/dall-e-3": { + "input_cost_per_pixel": 6.539e-08, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_token": 0.0 + }, + "azure/hd/1792-x-1024/dall-e-3": { + "input_cost_per_pixel": 6.539e-08, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_token": 0.0 + }, + "azure/high/1024-x-1024/gpt-image-1": { + "input_cost_per_pixel": 1.59263611e-07, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_pixel": 0.0, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "azure/high/1024-x-1536/gpt-image-1": { + "input_cost_per_pixel": 1.58945719e-07, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_pixel": 0.0, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "azure/high/1536-x-1024/gpt-image-1": { + "input_cost_per_pixel": 1.58945719e-07, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_pixel": 0.0, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "azure/low/1024-x-1024/gpt-image-1": { + "input_cost_per_pixel": 1.0490417e-08, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_pixel": 0.0, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "azure/low/1024-x-1536/gpt-image-1": { + "input_cost_per_pixel": 1.0172526e-08, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_pixel": 0.0, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "azure/low/1536-x-1024/gpt-image-1": { + "input_cost_per_pixel": 1.0172526e-08, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_pixel": 0.0, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "azure/medium/1024-x-1024/gpt-image-1": { + "input_cost_per_pixel": 4.0054321e-08, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_pixel": 0.0, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "azure/medium/1024-x-1536/gpt-image-1": { + "input_cost_per_pixel": 4.0054321e-08, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_pixel": 0.0, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "azure/medium/1536-x-1024/gpt-image-1": { + "input_cost_per_pixel": 4.0054321e-08, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_pixel": 0.0, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, "azure/gpt-image-1-mini": { "cache_read_input_image_token_cost": 2.5e-07, "cache_read_input_token_cost": 2e-07, @@ -4044,183 +4613,6 @@ "/v1/images/edits" ] }, - "azure/gpt-realtime-1.5-2026-02-23": { - "cache_creation_input_audio_token_cost": 4e-06, - "cache_read_input_token_cost": 4e-06, - "input_cost_per_audio_token": 3.2e-05, - "input_cost_per_image": 5e-06, - "input_cost_per_token": 4e-06, - "litellm_provider": "azure", - "max_input_tokens": 32000, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_audio_token": 6.4e-05, - "output_cost_per_token": 1.6e-05, - "supported_endpoints": [ - "/v1/realtime" - ], - "supported_modalities": [ - "text", - "image", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "azure/gpt-realtime-2025-08-28": { - "cache_creation_input_audio_token_cost": 4e-06, - "cache_read_input_token_cost": 4e-06, - "input_cost_per_audio_token": 3.2e-05, - "input_cost_per_image": 5e-06, - "input_cost_per_token": 4e-06, - "litellm_provider": "azure", - "max_input_tokens": 32000, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_audio_token": 6.4e-05, - "output_cost_per_token": 1.6e-05, - "supported_endpoints": [ - "/v1/realtime" - ], - "supported_modalities": [ - "text", - "image", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "azure/gpt-realtime-mini-2025-10-06": { - "cache_creation_input_audio_token_cost": 3e-07, - "cache_read_input_token_cost": 6e-08, - "input_cost_per_audio_token": 1e-05, - "input_cost_per_image": 8e-07, - "input_cost_per_token": 6e-07, - "litellm_provider": "azure", - "max_input_tokens": 32000, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_audio_token": 2e-05, - "output_cost_per_token": 2.4e-06, - "supported_endpoints": [ - "/v1/realtime" - ], - "supported_modalities": [ - "text", - "image", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "azure/hd/1024-x-1024/dall-e-3": { - "input_cost_per_pixel": 7.629e-08, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_token": 0.0 - }, - "azure/hd/1024-x-1792/dall-e-3": { - "input_cost_per_pixel": 6.539e-08, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_token": 0.0 - }, - "azure/hd/1792-x-1024/dall-e-3": { - "input_cost_per_pixel": 6.539e-08, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_token": 0.0 - }, - "azure/high/1024-x-1024/gpt-image-1": { - "input_cost_per_pixel": 1.59263611e-07, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_pixel": 0.0, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, - "azure/high/1024-x-1024/gpt-image-1-mini": { - "input_cost_per_pixel": 3.173828125e-08, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_pixel": 0.0, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, - "azure/high/1024-x-1536/gpt-image-1": { - "input_cost_per_pixel": 1.58945719e-07, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_pixel": 0.0, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, - "azure/high/1024-x-1536/gpt-image-1-mini": { - "input_cost_per_pixel": 3.173828125e-08, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_pixel": 0.0, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, - "azure/high/1536-x-1024/gpt-image-1": { - "input_cost_per_pixel": 1.58945719e-07, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_pixel": 0.0, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, - "azure/high/1536-x-1024/gpt-image-1-mini": { - "input_cost_per_pixel": 3.1575520833e-08, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_pixel": 0.0, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, - "azure/low/1024-x-1024/gpt-image-1": { - "input_cost_per_pixel": 1.0490417e-08, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_pixel": 0.0, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, "azure/low/1024-x-1024/gpt-image-1-mini": { "input_cost_per_pixel": 2.0751953125e-09, "litellm_provider": "azure", @@ -4230,15 +4622,6 @@ "/v1/images/generations" ] }, - "azure/low/1024-x-1536/gpt-image-1": { - "input_cost_per_pixel": 1.0172526e-08, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_pixel": 0.0, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, "azure/low/1024-x-1536/gpt-image-1-mini": { "input_cost_per_pixel": 2.0751953125e-09, "litellm_provider": "azure", @@ -4248,15 +4631,6 @@ "/v1/images/generations" ] }, - "azure/low/1536-x-1024/gpt-image-1": { - "input_cost_per_pixel": 1.0172526e-08, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_pixel": 0.0, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, "azure/low/1536-x-1024/gpt-image-1-mini": { "input_cost_per_pixel": 2.0345052083e-09, "litellm_provider": "azure", @@ -4266,15 +4640,6 @@ "/v1/images/generations" ] }, - "azure/medium/1024-x-1024/gpt-image-1": { - "input_cost_per_pixel": 4.0054321e-08, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_pixel": 0.0, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, "azure/medium/1024-x-1024/gpt-image-1-mini": { "input_cost_per_pixel": 8.056640625e-09, "litellm_provider": "azure", @@ -4284,15 +4649,6 @@ "/v1/images/generations" ] }, - "azure/medium/1024-x-1536/gpt-image-1": { - "input_cost_per_pixel": 4.0054321e-08, - "litellm_provider": "azure", - "mode": "image_generation", - "output_cost_per_pixel": 0.0, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, "azure/medium/1024-x-1536/gpt-image-1-mini": { "input_cost_per_pixel": 8.056640625e-09, "litellm_provider": "azure", @@ -4302,8 +4658,8 @@ "/v1/images/generations" ] }, - "azure/medium/1536-x-1024/gpt-image-1": { - "input_cost_per_pixel": 4.0054321e-08, + "azure/medium/1536-x-1024/gpt-image-1-mini": { + "input_cost_per_pixel": 7.9752604167e-09, "litellm_provider": "azure", "mode": "image_generation", "output_cost_per_pixel": 0.0, @@ -4311,8 +4667,26 @@ "/v1/images/generations" ] }, - "azure/medium/1536-x-1024/gpt-image-1-mini": { - "input_cost_per_pixel": 7.9752604167e-09, + "azure/high/1024-x-1024/gpt-image-1-mini": { + "input_cost_per_pixel": 3.173828125e-08, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_pixel": 0.0, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "azure/high/1024-x-1536/gpt-image-1-mini": { + "input_cost_per_pixel": 3.173828125e-08, + "litellm_provider": "azure", + "mode": "image_generation", + "output_cost_per_pixel": 0.0, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "azure/high/1536-x-1024/gpt-image-1-mini": { + "input_cost_per_pixel": 3.1575520833e-08, "litellm_provider": "azure", "mode": "image_generation", "output_cost_per_pixel": 0.0, @@ -4657,66 +5031,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "azure/sora-2": { - "litellm_provider": "azure", - "mode": "video_generation", - "output_cost_per_video_per_second": 0.1, - "source": "https://azure.microsoft.com/en-us/products/ai-services/video-generation", - "supported_modalities": [ - "text" - ], - "supported_output_modalities": [ - "video" - ], - "supported_resolutions": [ - "720x1280", - "1280x720" - ] - }, - "azure/sora-2-pro": { - "litellm_provider": "azure", - "mode": "video_generation", - "output_cost_per_video_per_second": 0.3, - "source": "https://azure.microsoft.com/en-us/products/ai-services/video-generation", - "supported_modalities": [ - "text" - ], - "supported_output_modalities": [ - "video" - ], - "supported_resolutions": [ - "720x1280", - "1280x720" - ] - }, - "azure/sora-2-pro-high-res": { - "litellm_provider": "azure", - "mode": "video_generation", - "output_cost_per_video_per_second": 0.5, - "source": "https://azure.microsoft.com/en-us/products/ai-services/video-generation", - "supported_modalities": [ - "text" - ], - "supported_output_modalities": [ - "video" - ], - "supported_resolutions": [ - "1024x1792", - "1792x1024" - ] - }, - "azure/speech/azure-tts": { - "input_cost_per_character": 1.5e-05, - "litellm_provider": "azure", - "mode": "audio_speech", - "source": "https://azure.microsoft.com/en-us/pricing/calculator/" - }, - "azure/speech/azure-tts-hd": { - "input_cost_per_character": 3e-05, - "litellm_provider": "azure", - "mode": "audio_speech", - "source": "https://azure.microsoft.com/en-us/pricing/calculator/" - }, "azure/standard/1024-x-1024/dall-e-2": { "input_cost_per_pixel": 0.0, "litellm_provider": "azure", @@ -4766,6 +5080,18 @@ "mode": "embedding", "output_cost_per_token": 0.0 }, + "azure/speech/azure-tts": { + "input_cost_per_character": 1.5e-05, + "litellm_provider": "azure", + "mode": "audio_speech", + "source": "https://azure.microsoft.com/en-us/pricing/calculator/" + }, + "azure/speech/azure-tts-hd": { + "input_cost_per_character": 3e-05, + "litellm_provider": "azure", + "mode": "audio_speech", + "source": "https://azure.microsoft.com/en-us/pricing/calculator/" + }, "azure/tts-1": { "input_cost_per_character": 1.5e-05, "litellm_provider": "azure", @@ -5369,6 +5695,15 @@ "/v1/images/generations" ] }, + "azure_ai/flux.2-pro": { + "litellm_provider": "azure_ai", + "mode": "image_generation", + "output_cost_per_image": 0.04, + "source": "https://ai.azure.com/explore/models/flux.2-pro/version/1/registry/azureml-blackforestlabs", + "supported_endpoints": [ + "/v1/images/generations" + ] + }, "azure_ai/Llama-3.2-11B-Vision-Instruct": { "input_cost_per_token": 3.7e-07, "litellm_provider": "azure_ai", @@ -5433,18 +5768,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "azure_ai/MAI-DS-R1": { - "input_cost_per_token": 1.35e-06, - "litellm_provider": "azure_ai", - "max_input_tokens": 128000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 5.4e-06, - "source": "https://azure.microsoft.com/en-us/pricing/details/ai-foundry-models/microsoft/", - "supports_reasoning": true, - "supports_tool_choice": true - }, "azure_ai/Meta-Llama-3-70B-Instruct": { "input_cost_per_token": 1.1e-06, "litellm_provider": "azure_ai", @@ -5620,17 +5943,6 @@ "source": "https://techcommunity.microsoft.com/blog/Azure-AI-Services-blog/announcing-new-phi-pricing-empowering-your-business-with-small-language-models/4395112", "supports_function_calling": true }, - "azure_ai/Phi-4-mini-reasoning": { - "input_cost_per_token": 8e-08, - "litellm_provider": "azure_ai", - "max_input_tokens": 131072, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_token": 3.2e-07, - "source": "https://azure.microsoft.com/en-us/pricing/details/ai-foundry-models/microsoft/", - "supports_function_calling": true - }, "azure_ai/Phi-4-multimodal-instruct": { "input_cost_per_audio_token": 4e-06, "input_cost_per_token": 8e-08, @@ -5645,6 +5957,17 @@ "supports_function_calling": true, "supports_vision": true }, + "azure_ai/Phi-4-mini-reasoning": { + "input_cost_per_token": 8e-08, + "litellm_provider": "azure_ai", + "max_input_tokens": 131072, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_token": 3.2e-07, + "source": "https://azure.microsoft.com/en-us/pricing/details/ai-foundry-models/microsoft/", + "supports_function_calling": true + }, "azure_ai/Phi-4-reasoning": { "input_cost_per_token": 1.25e-07, "litellm_provider": "azure_ai", @@ -5658,138 +5981,62 @@ "supports_tool_choice": true, "supports_reasoning": true }, - "azure_ai/claude-haiku-4-5": { - "cache_creation_input_token_cost": 1.25e-06, - "cache_creation_input_token_cost_above_1hr": 2e-06, - "cache_read_input_token_cost": 1e-07, - "input_cost_per_token": 1e-06, + "azure_ai/mistral-document-ai-2505": { "litellm_provider": "azure_ai", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 5e-06, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true + "ocr_cost_per_page": 0.003, + "mode": "ocr", + "supported_endpoints": [ + "/v1/ocr" + ], + "source": "https://devblogs.microsoft.com/foundry/whats-new-in-azure-ai-foundry-august-2025/#mistral-document-ai-(ocr)-%E2%80%94-serverless-in-foundry" }, - "azure_ai/claude-opus-4-1": { - "cache_creation_input_token_cost": 1.875e-05, - "cache_creation_input_token_cost_above_1hr": 3e-05, - "cache_read_input_token_cost": 1.5e-06, - "input_cost_per_token": 1.5e-05, + "azure_ai/mistral-document-ai-2512": { "litellm_provider": "azure_ai", - "max_input_tokens": 200000, - "max_output_tokens": 32000, - "max_tokens": 32000, - "mode": "chat", - "output_cost_per_token": 7.5e-05, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true + "ocr_cost_per_page": 0.003, + "mode": "ocr", + "supported_endpoints": [ + "/v1/ocr" + ], + "source": "https://azure.microsoft.com/en-us/pricing/details/ai-foundry-models/" }, - "azure_ai/claude-opus-4-5": { - "cache_creation_input_token_cost": 6.25e-06, - "cache_creation_input_token_cost_above_1hr": 1e-05, - "cache_read_input_token_cost": 5e-07, - "input_cost_per_token": 5e-06, + "azure_ai/doc-intelligence/prebuilt-read": { "litellm_provider": "azure_ai", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 2.5e-05, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true + "ocr_cost_per_page": 0.0015, + "mode": "ocr", + "supported_endpoints": [ + "/v1/ocr" + ], + "source": "https://azure.microsoft.com/en-us/pricing/details/ai-document-intelligence/" }, - "azure_ai/claude-opus-4-6": { - "input_cost_per_token": 5e-06, - "output_cost_per_token": 2.5e-05, + "azure_ai/doc-intelligence/prebuilt-layout": { "litellm_provider": "azure_ai", - "max_input_tokens": 200000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "cache_creation_input_token_cost": 6.25e-06, - "cache_creation_input_token_cost_above_1hr": 1e-05, - "cache_read_input_token_cost": 5e-07, - "supports_assistant_prefill": false, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 159 + "ocr_cost_per_page": 0.01, + "mode": "ocr", + "supported_endpoints": [ + "/v1/ocr" + ], + "source": "https://azure.microsoft.com/en-us/pricing/details/ai-document-intelligence/" }, - "azure_ai/claude-sonnet-4-5": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_creation_input_token_cost_above_1hr": 6e-06, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 3e-06, + "azure_ai/doc-intelligence/prebuilt-document": { "litellm_provider": "azure_ai", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true + "ocr_cost_per_page": 0.01, + "mode": "ocr", + "supported_endpoints": [ + "/v1/ocr" + ], + "source": "https://azure.microsoft.com/en-us/pricing/details/ai-document-intelligence/" }, - "azure_ai/claude-sonnet-4-6": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_creation_input_token_cost_above_1hr": 6e-06, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 3e-06, + "azure_ai/MAI-DS-R1": { + "input_cost_per_token": 1.35e-06, "litellm_provider": "azure_ai", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, + "max_input_tokens": 128000, + "max_output_tokens": 8192, + "max_tokens": 8192, "mode": "chat", - "output_cost_per_token": 1.5e-05, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, + "output_cost_per_token": 5.4e-06, + "source": "https://azure.microsoft.com/en-us/pricing/details/ai-foundry-models/microsoft/", "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 + "supports_tool_choice": true }, "azure_ai/cohere-rerank-v3-english": { "input_cost_per_query": 0.002, @@ -5824,6 +6071,17 @@ "mode": "rerank", "output_cost_per_token": 0.0 }, + "azure_ai/cohere-rerank-v4.0-pro": { + "input_cost_per_query": 0.0025, + "input_cost_per_token": 0.0, + "litellm_provider": "azure_ai", + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "max_query_tokens": 4096, + "max_tokens": 32768, + "mode": "rerank", + "output_cost_per_token": 0.0 + }, "azure_ai/cohere-rerank-v4.0-fast": { "input_cost_per_query": 0.002, "input_cost_per_token": 0.0, @@ -5835,16 +6093,33 @@ "mode": "rerank", "output_cost_per_token": 0.0 }, - "azure_ai/cohere-rerank-v4.0-pro": { - "input_cost_per_query": 0.0025, - "input_cost_per_token": 0.0, + "azure_ai/deepseek-v3.2": { + "input_cost_per_token": 5.8e-07, "litellm_provider": "azure_ai", - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "max_query_tokens": 4096, - "max_tokens": 32768, - "mode": "rerank", - "output_cost_per_token": 0.0 + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "max_tokens": 163840, + "mode": "chat", + "output_cost_per_token": 1.68e-06, + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, + "azure_ai/deepseek-v3.2-speciale": { + "input_cost_per_token": 5.8e-07, + "litellm_provider": "azure_ai", + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "max_tokens": 163840, + "mode": "chat", + "output_cost_per_token": 1.68e-06, + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true }, "azure_ai/deepseek-r1": { "input_cost_per_token": 1.35e-06, @@ -5881,61 +6156,6 @@ "supports_function_calling": true, "supports_tool_choice": true }, - "azure_ai/deepseek-v3.2": { - "input_cost_per_token": 5.8e-07, - "litellm_provider": "azure_ai", - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "max_tokens": 163840, - "mode": "chat", - "output_cost_per_token": 1.68e-06, - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, - "azure_ai/deepseek-v3.2-speciale": { - "input_cost_per_token": 5.8e-07, - "litellm_provider": "azure_ai", - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "max_tokens": 163840, - "mode": "chat", - "output_cost_per_token": 1.68e-06, - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, - "azure_ai/doc-intelligence/prebuilt-document": { - "litellm_provider": "azure_ai", - "ocr_cost_per_page": 0.01, - "mode": "ocr", - "supported_endpoints": [ - "/v1/ocr" - ], - "source": "https://azure.microsoft.com/en-us/pricing/details/ai-document-intelligence/" - }, - "azure_ai/doc-intelligence/prebuilt-layout": { - "litellm_provider": "azure_ai", - "ocr_cost_per_page": 0.01, - "mode": "ocr", - "supported_endpoints": [ - "/v1/ocr" - ], - "source": "https://azure.microsoft.com/en-us/pricing/details/ai-document-intelligence/" - }, - "azure_ai/doc-intelligence/prebuilt-read": { - "litellm_provider": "azure_ai", - "ocr_cost_per_page": 0.0015, - "mode": "ocr", - "supported_endpoints": [ - "/v1/ocr" - ], - "source": "https://azure.microsoft.com/en-us/pricing/details/ai-document-intelligence/" - }, "azure_ai/embed-v-4-0": { "input_cost_per_token": 1.2e-07, "litellm_provider": "azure_ai", @@ -5954,15 +6174,6 @@ ], "supports_embedding_image_input": true }, - "azure_ai/flux.2-pro": { - "litellm_provider": "azure_ai", - "mode": "image_generation", - "output_cost_per_image": 0.04, - "source": "https://ai.azure.com/explore/models/flux.2-pro/version/1/registry/azureml-blackforestlabs", - "supported_endpoints": [ - "/v1/images/generations" - ] - }, "azure_ai/global/grok-3": { "input_cost_per_token": 3e-06, "litellm_provider": "azure_ai", @@ -5992,20 +6203,6 @@ "supports_tool_choice": true, "supports_web_search": true }, - "azure_ai/gpt-oss-120b": { - "input_cost_per_token": 1.5e-07, - "output_cost_per_token": 6e-07, - "litellm_provider": "azure_ai", - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "max_tokens": 131072, - "mode": "chat", - "source": "https://azure.microsoft.com/en-us/pricing/details/cognitive-services/openai-service/", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, "azure_ai/grok-3": { "input_cost_per_token": 3e-06, "litellm_provider": "azure_ai", @@ -6049,6 +6246,33 @@ "supports_tool_choice": true, "supports_web_search": true }, + "azure_ai/grok-4-fast-non-reasoning": { + "input_cost_per_token": 2e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "azure_ai", + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "max_tokens": 131072, + "mode": "chat", + "supports_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_web_search": true + }, + "azure_ai/grok-4-fast-reasoning": { + "input_cost_per_token": 2e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "azure_ai", + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "max_tokens": 131072, + "mode": "chat", + "source": "https://azure.microsoft.com/en-us/pricing/details/ai-foundry-models/grok/", + "supports_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_web_search": true + }, "azure_ai/grok-4-1-fast-non-reasoning": { "input_cost_per_token": 2e-07, "output_cost_per_token": 5e-07, @@ -6078,33 +6302,6 @@ "supports_tool_choice": true, "supports_web_search": true }, - "azure_ai/grok-4-fast-non-reasoning": { - "input_cost_per_token": 2e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "azure_ai", - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "max_tokens": 131072, - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_web_search": true - }, - "azure_ai/grok-4-fast-reasoning": { - "input_cost_per_token": 2e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "azure_ai", - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "max_tokens": 131072, - "mode": "chat", - "source": "https://azure.microsoft.com/en-us/pricing/details/ai-foundry-models/grok/", - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_web_search": true - }, "azure_ai/grok-code-fast-1": { "input_cost_per_token": 2e-07, "litellm_provider": "azure_ai", @@ -6165,24 +6362,6 @@ "supports_function_calling": true, "supports_tool_choice": true }, - "azure_ai/mistral-document-ai-2505": { - "litellm_provider": "azure_ai", - "ocr_cost_per_page": 0.003, - "mode": "ocr", - "supported_endpoints": [ - "/v1/ocr" - ], - "source": "https://devblogs.microsoft.com/foundry/whats-new-in-azure-ai-foundry-august-2025/#mistral-document-ai-(ocr)-%E2%80%94-serverless-in-foundry" - }, - "azure_ai/mistral-document-ai-2512": { - "litellm_provider": "azure_ai", - "ocr_cost_per_page": 0.003, - "mode": "ocr", - "supported_endpoints": [ - "/v1/ocr" - ], - "source": "https://azure.microsoft.com/en-us/pricing/details/ai-foundry-models/" - }, "azure_ai/mistral-large": { "input_cost_per_token": 4e-06, "litellm_provider": "azure_ai", @@ -6206,6 +6385,18 @@ "supports_function_calling": true, "supports_tool_choice": true }, + "azure_ai/mistral-large-latest": { + "input_cost_per_token": 2e-06, + "litellm_provider": "azure_ai", + "max_input_tokens": 128000, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_token": 6e-06, + "source": "https://azuremarketplace.microsoft.com/en/marketplace/apps/000-000.mistral-ai-large-2407-offer?tab=Overview", + "supports_function_calling": true, + "supports_tool_choice": true + }, "azure_ai/mistral-large-3": { "input_cost_per_token": 5e-07, "litellm_provider": "azure_ai", @@ -6219,18 +6410,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "azure_ai/mistral-large-latest": { - "input_cost_per_token": 2e-06, - "litellm_provider": "azure_ai", - "max_input_tokens": 128000, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_token": 6e-06, - "source": "https://azuremarketplace.microsoft.com/en/marketplace/apps/000-000.mistral-ai-large-2407-offer?tab=Overview", - "supports_function_calling": true, - "supports_tool_choice": true - }, "azure_ai/mistral-medium-2505": { "input_cost_per_token": 4e-07, "litellm_provider": "azure_ai", @@ -6277,14 +6456,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "azure_ai/model_router": { - "input_cost_per_token": 1.4e-07, - "output_cost_per_token": 0, - "litellm_provider": "azure_ai", - "mode": "chat", - "source": "https://azure.microsoft.com/en-us/pricing/details/ai-services/", - "comment": "Flat cost of $0.14 per M input tokens for Azure AI Foundry Model Router infrastructure. Use pattern: azure_ai/model_router/ where deployment-name is your Azure deployment (e.g., azure-model-router)" - }, "babbage-002": { "input_cost_per_token": 4e-07, "litellm_provider": "text-completion-openai", @@ -6486,18 +6657,32 @@ "supports_tool_choice": true, "source": "https://aws.amazon.com/bedrock/pricing/" }, - "bedrock/ap-south-1/deepseek.v3.2": { - "input_cost_per_token": 7.4e-07, + "bedrock/moonshotai.kimi-k2-thinking": { + "input_cost_per_token": 7.3e-07, "litellm_provider": "bedrock", - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "max_tokens": 163840, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, "mode": "chat", - "output_cost_per_token": 2.22e-06, + "output_cost_per_token": 3.03e-06, + "supports_function_calling": true, + "supports_reasoning": true + }, + "bedrock/moonshotai.kimi-k2.5": { + "input_cost_per_token": 6e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 3.03e-06, + "source": "https://platform.moonshot.ai/docs/guide/kimi-k2-5-quickstart", "supports_function_calling": true, "supports_reasoning": true, + "supports_system_messages": true, "supports_tool_choice": true, - "source": "https://aws.amazon.com/bedrock/pricing/" + "supports_video_input": true, + "supports_vision": true }, "bedrock/ap-south-1/meta.llama3-70b-instruct-v1:0": { "input_cost_per_token": 3.18e-06, @@ -6517,6 +6702,19 @@ "mode": "chat", "output_cost_per_token": 7.2e-07 }, + "bedrock/ap-south-1/deepseek.v3.2": { + "input_cost_per_token": 7.4e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "max_tokens": 163840, + "mode": "chat", + "output_cost_per_token": 2.22e-06, + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, "bedrock/ap-south-1/minimax.minimax-m2.1": { "input_cost_per_token": 3.6e-07, "litellm_provider": "bedrock", @@ -6639,6 +6837,46 @@ "mode": "chat", "output_cost_per_token": 6.9e-07 }, + "bedrock/eu-north-1/deepseek.v3.2": { + "input_cost_per_token": 7.4e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "max_tokens": 163840, + "mode": "chat", + "output_cost_per_token": 2.22e-06, + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, + "bedrock/eu-north-1/minimax.minimax-m2.1": { + "input_cost_per_token": 3.6e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 196000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 1.44e-06, + "supports_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, + "bedrock/eu-north-1/moonshotai.kimi-k2.5": { + "input_cost_per_token": 7.2e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 3.6e-06, + "supports_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, "bedrock/eu-central-1/1-month-commitment/anthropic.claude-instant-v1": { "input_cost_per_second": 0.01635, "litellm_provider": "bedrock", @@ -6752,72 +6990,6 @@ "supports_tool_choice": true, "source": "https://aws.amazon.com/bedrock/pricing/" }, - "bedrock/eu-north-1/deepseek.v3.2": { - "input_cost_per_token": 7.4e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "max_tokens": 163840, - "mode": "chat", - "output_cost_per_token": 2.22e-06, - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, - "bedrock/eu-north-1/minimax.minimax-m2.1": { - "input_cost_per_token": 3.6e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 196000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1.44e-06, - "supports_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, - "bedrock/eu-north-1/moonshotai.kimi-k2.5": { - "input_cost_per_token": 7.2e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 3.6e-06, - "supports_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, - "bedrock/eu-south-1/minimax.minimax-m2.1": { - "input_cost_per_token": 3.6e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 196000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1.44e-06, - "supports_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, - "bedrock/eu-south-1/qwen.qwen3-coder-next": { - "input_cost_per_token": 6e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 262144, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1.44e-06, - "supports_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, "bedrock/eu-west-1/meta.llama3-70b-instruct-v1:0": { "input_cost_per_token": 2.86e-06, "litellm_provider": "bedrock", @@ -6936,6 +7108,32 @@ "output_cost_per_token": 9.1e-07, "supports_tool_choice": true }, + "bedrock/eu-south-1/minimax.minimax-m2.1": { + "input_cost_per_token": 3.6e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 196000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 1.44e-06, + "supports_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, + "bedrock/eu-south-1/qwen.qwen3-coder-next": { + "input_cost_per_token": 6e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 262144, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 1.44e-06, + "supports_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, "bedrock/invoke/anthropic.claude-3-5-sonnet-20240620-v1:0": { "input_cost_per_token": 3e-06, "litellm_provider": "bedrock", @@ -6954,46 +7152,6 @@ "cache_read_input_token_cost": 3e-07, "cache_creation_input_token_cost": 3.75e-06 }, - "bedrock/moonshotai.kimi-k2-thinking": { - "input_cost_per_token": 7.3e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 3.03e-06, - "supports_function_calling": true, - "supports_reasoning": true - }, - "bedrock/moonshotai.kimi-k2.5": { - "input_cost_per_token": 6e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 3.03e-06, - "source": "https://platform.moonshot.ai/docs/guide/kimi-k2-5-quickstart", - "supports_function_calling": true, - "supports_reasoning": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_video_input": true, - "supports_vision": true - }, - "bedrock/sa-east-1/deepseek.v3.2": { - "input_cost_per_token": 7.4e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "max_tokens": 163840, - "mode": "chat", - "output_cost_per_token": 2.22e-06, - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, "bedrock/sa-east-1/meta.llama3-70b-instruct-v1:0": { "input_cost_per_token": 4.45e-06, "litellm_provider": "bedrock", @@ -7012,6 +7170,19 @@ "mode": "chat", "output_cost_per_token": 1.01e-06 }, + "bedrock/sa-east-1/deepseek.v3.2": { + "input_cost_per_token": 7.4e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "max_tokens": 163840, + "mode": "chat", + "output_cost_per_token": 2.22e-06, + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, "bedrock/sa-east-1/minimax.minimax-m2.1": { "input_cost_per_token": 3.6e-07, "litellm_provider": "bedrock", @@ -7151,19 +7322,6 @@ "output_cost_per_token": 2.4e-05, "supports_tool_choice": true }, - "bedrock/us-east-1/deepseek.v3.2": { - "input_cost_per_token": 6.2e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "max_tokens": 163840, - "mode": "chat", - "output_cost_per_token": 1.85e-06, - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, "bedrock/us-east-1/meta.llama3-70b-instruct-v1:0": { "input_cost_per_token": 2.65e-06, "litellm_provider": "bedrock", @@ -7182,19 +7340,6 @@ "mode": "chat", "output_cost_per_token": 6e-07 }, - "bedrock/us-east-1/minimax.minimax-m2.1": { - "input_cost_per_token": 3e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 196000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1.2e-06, - "supports_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, "bedrock/us-east-1/mistral.mistral-7b-instruct-v0:2": { "input_cost_per_token": 1.5e-07, "litellm_provider": "bedrock", @@ -7225,6 +7370,32 @@ "output_cost_per_token": 7e-07, "supports_tool_choice": true }, + "bedrock/us-east-1/deepseek.v3.2": { + "input_cost_per_token": 6.2e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "max_tokens": 163840, + "mode": "chat", + "output_cost_per_token": 1.85e-06, + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, + "bedrock/us-east-1/minimax.minimax-m2.1": { + "input_cost_per_token": 3e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 196000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 1.2e-06, + "supports_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, "bedrock/us-east-1/moonshotai.kimi-k2-thinking": { "input_cost_per_token": 6e-07, "litellm_provider": "bedrock", @@ -7517,22 +7688,6 @@ "mode": "chat", "output_cost_per_token": 1.5e-06 }, - "bedrock/us-gov-west-1/anthropic.claude-3-5-sonnet-20240620-v1:0": { - "input_cost_per_token": 3.6e-06, - "litellm_provider": "bedrock", - "max_input_tokens": 200000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1.8e-05, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "cache_read_input_token_cost": 3.6e-07, - "cache_creation_input_token_cost": 4.5e-06 - }, "bedrock/us-gov-west-1/anthropic.claude-3-7-sonnet-20250219-v1:0": { "cache_creation_input_token_cost": 4.5e-06, "cache_read_input_token_cost": 3.6e-07, @@ -7553,6 +7708,22 @@ "supports_tool_choice": true, "supports_vision": true }, + "bedrock/us-gov-west-1/anthropic.claude-3-5-sonnet-20240620-v1:0": { + "input_cost_per_token": 3.6e-06, + "litellm_provider": "bedrock", + "max_input_tokens": 200000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 1.8e-05, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "cache_read_input_token_cost": 3.6e-07, + "cache_creation_input_token_cost": 4.5e-06 + }, "bedrock/us-gov-west-1/anthropic.claude-3-haiku-20240307-v1:0": { "input_cost_per_token": 3e-07, "litellm_provider": "bedrock", @@ -7715,32 +7886,6 @@ "output_cost_per_token": 2.4e-05, "supports_tool_choice": true }, - "bedrock/us-west-2/deepseek.v3.2": { - "input_cost_per_token": 6.2e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "max_tokens": 163840, - "mode": "chat", - "output_cost_per_token": 1.85e-06, - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, - "bedrock/us-west-2/minimax.minimax-m2.1": { - "input_cost_per_token": 3e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 196000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1.2e-06, - "supports_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, "bedrock/us-west-2/mistral.mistral-7b-instruct-v0:2": { "input_cost_per_token": 1.5e-07, "litellm_provider": "bedrock", @@ -7771,6 +7916,32 @@ "output_cost_per_token": 7e-07, "supports_tool_choice": true }, + "bedrock/us-west-2/deepseek.v3.2": { + "input_cost_per_token": 6.2e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "max_tokens": 163840, + "mode": "chat", + "output_cost_per_token": 1.85e-06, + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, + "bedrock/us-west-2/minimax.minimax-m2.1": { + "input_cost_per_token": 3e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 196000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 1.2e-06, + "supports_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, "bedrock/us-west-2/moonshotai.kimi-k2-thinking": { "input_cost_per_token": 6e-07, "litellm_provider": "bedrock", @@ -7826,66 +7997,13 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "bedrock_mantle/openai.gpt-oss-120b": { - "input_cost_per_token": 1.5e-07, - "output_cost_per_token": 6e-07, - "litellm_provider": "bedrock_mantle", - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "chat", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "bedrock_mantle/openai.gpt-oss-20b": { - "input_cost_per_token": 7.5e-08, - "output_cost_per_token": 3e-07, - "litellm_provider": "bedrock_mantle", - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "chat", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "bedrock_mantle/openai.gpt-oss-safeguard-120b": { - "input_cost_per_token": 1.5e-07, - "output_cost_per_token": 6e-07, - "litellm_provider": "bedrock_mantle", - "max_input_tokens": 131072, - "max_output_tokens": 65536, - "max_tokens": 65536, - "mode": "chat", - "supports_function_calling": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "bedrock_mantle/openai.gpt-oss-safeguard-20b": { - "input_cost_per_token": 7.5e-08, - "output_cost_per_token": 3e-07, - "litellm_provider": "bedrock_mantle", - "max_input_tokens": 131072, - "max_output_tokens": 65536, - "max_tokens": 65536, - "mode": "chat", - "supports_function_calling": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "black_forest_labs/flux-dev": { + "black_forest_labs/flux-kontext-pro": { "litellm_provider": "black_forest_labs", - "mode": "image_generation", - "output_cost_per_image": 0.025, + "mode": "image_edit", + "output_cost_per_image": 0.04, "source": "https://bfl.ai/pricing", "supported_endpoints": [ + "/v1/images/edits", "/v1/images/generations" ] }, @@ -7899,26 +8017,7 @@ "/v1/images/generations" ] }, - "black_forest_labs/flux-kontext-pro": { - "litellm_provider": "black_forest_labs", - "mode": "image_edit", - "output_cost_per_image": 0.04, - "source": "https://bfl.ai/pricing", - "supported_endpoints": [ - "/v1/images/edits", - "/v1/images/generations" - ] - }, - "black_forest_labs/flux-pro": { - "litellm_provider": "black_forest_labs", - "mode": "image_generation", - "output_cost_per_image": 0.05, - "source": "https://bfl.ai/pricing", - "supported_endpoints": [ - "/v1/images/generations" - ] - }, - "black_forest_labs/flux-pro-1.0-expand": { + "black_forest_labs/flux-pro-1.0-fill": { "litellm_provider": "black_forest_labs", "mode": "image_edit", "output_cost_per_image": 0.05, @@ -7927,7 +8026,7 @@ "/v1/images/edits" ] }, - "black_forest_labs/flux-pro-1.0-fill": { + "black_forest_labs/flux-pro-1.0-expand": { "litellm_provider": "black_forest_labs", "mode": "image_edit", "output_cost_per_image": 0.05, @@ -7954,20 +8053,23 @@ "/v1/images/generations" ] }, - "cerebras/gpt-oss-120b": { - "input_cost_per_token": 3.5e-07, - "litellm_provider": "cerebras", - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "chat", - "output_cost_per_token": 7.5e-07, - "source": "https://www.cerebras.ai/blog/openai-gpt-oss-120b-runs-fastest-on-cerebras", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true + "black_forest_labs/flux-dev": { + "litellm_provider": "black_forest_labs", + "mode": "image_generation", + "output_cost_per_image": 0.025, + "source": "https://bfl.ai/pricing", + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "black_forest_labs/flux-pro": { + "litellm_provider": "black_forest_labs", + "mode": "image_generation", + "output_cost_per_image": 0.05, + "source": "https://bfl.ai/pricing", + "supported_endpoints": [ + "/v1/images/generations" + ] }, "cerebras/llama-3.3-70b": { "input_cost_per_token": 8.5e-07, @@ -8002,6 +8104,21 @@ "supports_function_calling": true, "supports_tool_choice": true }, + "cerebras/gpt-oss-120b": { + "input_cost_per_token": 3.5e-07, + "litellm_provider": "cerebras", + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "chat", + "output_cost_per_token": 7.5e-07, + "source": "https://www.cerebras.ai/blog/openai-gpt-oss-120b-runs-fastest-on-cerebras", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, "cerebras/qwen-3-32b": { "input_cost_per_token": 4e-07, "litellm_provider": "cerebras", @@ -8067,161 +8184,58 @@ "supports_tool_choice": true, "supports_vision": true }, - "chatgpt-image-latest": { - "cache_read_input_image_token_cost": 2.5e-06, - "cache_read_input_token_cost": 1.25e-06, - "input_cost_per_image_token": 1e-05, - "input_cost_per_token": 5e-06, + "gpt-4o-transcribe-diarize": { + "input_cost_per_audio_token": 6e-06, + "input_cost_per_token": 2.5e-06, "litellm_provider": "openai", - "mode": "image_generation", - "output_cost_per_image_token": 4e-05, + "max_input_tokens": 16000, + "max_output_tokens": 2000, + "mode": "audio_transcription", + "output_cost_per_token": 1e-05, "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" + "/v1/audio/transcriptions" ] }, - "chatgpt/gpt-5.1-codex-max": { - "litellm_provider": "chatgpt", - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "supported_endpoints": [ - "/v1/responses" - ], - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_vision": true - }, - "chatgpt/gpt-5.1-codex-mini": { - "litellm_provider": "chatgpt", - "max_input_tokens": 128000, + "claude-haiku-4-5-20251001": { + "cache_creation_input_token_cost": 1.25e-06, + "cache_creation_input_token_cost_above_1hr": 2e-06, + "cache_read_input_token_cost": 1e-07, + "input_cost_per_token": 1e-06, + "litellm_provider": "anthropic", + "max_input_tokens": 200000, "max_output_tokens": 64000, "max_tokens": 64000, - "mode": "responses", - "supported_endpoints": [ - "/v1/responses" - ], + "mode": "chat", + "output_cost_per_token": 5e-06, + "supports_assistant_prefill": true, "supports_function_calling": true, - "supports_parallel_function_calling": true, + "supports_computer_use": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, "supports_response_schema": true, + "supports_tool_choice": true, "supports_vision": true }, - "chatgpt/gpt-5.2": { - "litellm_provider": "chatgpt", - "max_input_tokens": 128000, + "claude-haiku-4-5": { + "cache_creation_input_token_cost": 1.25e-06, + "cache_creation_input_token_cost_above_1hr": 2e-06, + "cache_read_input_token_cost": 1e-07, + "input_cost_per_token": 1e-06, + "litellm_provider": "anthropic", + "max_input_tokens": 200000, "max_output_tokens": 64000, "max_tokens": 64000, - "mode": "responses", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/responses" - ], + "mode": "chat", + "output_cost_per_token": 5e-06, + "supports_assistant_prefill": true, "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_vision": true - }, - "chatgpt/gpt-5.2-codex": { - "litellm_provider": "chatgpt", - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "supported_endpoints": [ - "/v1/responses" - ], - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_vision": true - }, - "chatgpt/gpt-5.3-chat-latest": { - "litellm_provider": "chatgpt", - "max_input_tokens": 128000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "responses", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/responses" - ], - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_vision": true - }, - "chatgpt/gpt-5.3-codex": { - "litellm_provider": "chatgpt", - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "supported_endpoints": [ - "/v1/responses" - ], - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_vision": true - }, - "chatgpt/gpt-5.3-codex-spark": { - "litellm_provider": "chatgpt", - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "supported_endpoints": [ - "/v1/responses" - ], - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_vision": true - }, - "chatgpt/gpt-5.3-instant": { - "litellm_provider": "chatgpt", - "max_input_tokens": 128000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "responses", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/responses" - ], - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_vision": true - }, - "chatgpt/gpt-5.4": { - "litellm_provider": "chatgpt", - "max_input_tokens": 1050000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/responses" - ], - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_vision": true - }, - "chatgpt/gpt-5.4-pro": { - "litellm_provider": "chatgpt", - "max_input_tokens": 1050000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "supported_endpoints": [ - "/v1/responses" - ], - "supports_function_calling": true, - "supports_parallel_function_calling": true, + "supports_computer_use": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, "supports_response_schema": true, + "supports_tool_choice": true, "supports_vision": true }, "claude-3-7-sonnet-20250219": { @@ -8349,47 +8363,121 @@ "supports_web_search": true, "tool_use_system_prompt_tokens": 159 }, - "claude-haiku-4-5": { - "cache_creation_input_token_cost": 1.25e-06, - "cache_creation_input_token_cost_above_1hr": 2e-06, - "cache_read_input_token_cost": 1e-07, - "input_cost_per_token": 1e-06, + "claude-sonnet-4-5": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, "litellm_provider": "anthropic", "max_input_tokens": 200000, "max_output_tokens": 64000, "max_tokens": 64000, "mode": "chat", - "output_cost_per_token": 5e-06, + "output_cost_per_token": 1.5e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, "supports_assistant_prefill": true, - "supports_function_calling": true, "supports_computer_use": true, + "supports_function_calling": true, "supports_pdf_input": true, "supports_prompt_caching": true, "supports_reasoning": true, "supports_response_schema": true, "supports_tool_choice": true, - "supports_vision": true + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 }, - "claude-haiku-4-5-20251001": { - "cache_creation_input_token_cost": 1.25e-06, - "cache_creation_input_token_cost_above_1hr": 2e-06, - "cache_read_input_token_cost": 1e-07, - "input_cost_per_token": 1e-06, + "claude-sonnet-4-5-20250929": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, "litellm_provider": "anthropic", "max_input_tokens": 200000, "max_output_tokens": 64000, "max_tokens": 64000, "mode": "chat", - "output_cost_per_token": 5e-06, + "output_cost_per_token": 1.5e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, "supports_assistant_prefill": true, - "supports_function_calling": true, "supports_computer_use": true, + "supports_function_calling": true, "supports_pdf_input": true, "supports_prompt_caching": true, "supports_reasoning": true, "supports_response_schema": true, "supports_tool_choice": true, - "supports_vision": true + "supports_vision": true, + "supports_web_search": true, + "tool_use_system_prompt_tokens": 346 + }, + "claude-sonnet-4-6": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost": 3e-07, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "litellm_provider": "anthropic", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "claude-sonnet-4-5-20250929-v1:0": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, + "litellm_provider": "bedrock", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 159 }, "claude-opus-4-1": { "cache_creation_input_token_cost": 1.875e-05, @@ -8474,7 +8562,7 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, - "claude-opus-4-5": { + "claude-opus-4-5-20251101": { "cache_creation_input_token_cost": 6.25e-06, "cache_creation_input_token_cost_above_1hr": 1e-05, "cache_read_input_token_cost": 5e-07, @@ -8501,7 +8589,7 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, - "claude-opus-4-5-20251101": { + "claude-opus-4-5": { "cache_creation_input_token_cost": 6.25e-06, "cache_creation_input_token_cost_above_1hr": 1e-05, "cache_read_input_token_cost": 5e-07, @@ -8630,122 +8718,6 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, - "claude-sonnet-4-5": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "litellm_provider": "anthropic", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, - "claude-sonnet-4-5-20250929": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "litellm_provider": "anthropic", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "tool_use_system_prompt_tokens": 346 - }, - "claude-sonnet-4-5-20250929-v1:0": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "litellm_provider": "bedrock", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 159 - }, - "claude-sonnet-4-6": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost": 3e-07, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "litellm_provider": "anthropic", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, "cloudflare/@cf/meta/llama-2-7b-chat-fp16": { "input_cost_per_token": 1.923e-06, "litellm_provider": "cloudflare", @@ -8903,6 +8875,16 @@ "output_vector_size": 1536, "supports_embedding_image_input": true }, + "cohere/embed-v4.0": { + "input_cost_per_token": 1.2e-07, + "litellm_provider": "cohere", + "max_input_tokens": 128000, + "max_tokens": 128000, + "mode": "embedding", + "output_cost_per_token": 0.0, + "output_vector_size": 1536, + "supports_embedding_image_input": true + }, "cohere.rerank-v3-5:0": { "input_cost_per_query": 0.002, "input_cost_per_token": 0.0, @@ -8916,16 +8898,6 @@ "mode": "rerank", "output_cost_per_token": 0.0 }, - "cohere/embed-v4.0": { - "input_cost_per_token": 1.2e-07, - "litellm_provider": "cohere", - "max_input_tokens": 128000, - "max_tokens": 128000, - "mode": "embedding", - "output_cost_per_token": 0.0, - "output_vector_size": 1536, - "supports_embedding_image_input": true - }, "command": { "input_cost_per_token": 1e-06, "litellm_provider": "cohere", @@ -9066,6 +9038,49 @@ "/v1/images/generations" ] }, + "deepseek-chat": { + "cache_read_input_token_cost": 2.8e-08, + "input_cost_per_token": 2.8e-07, + "litellm_provider": "deepseek", + "max_input_tokens": 131072, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 4.2e-07, + "source": "https://api-docs.deepseek.com/quick_start/pricing", + "supported_endpoints": [ + "/v1/chat/completions" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "deepseek-reasoner": { + "cache_read_input_token_cost": 2.8e-08, + "input_cost_per_token": 2.8e-07, + "litellm_provider": "deepseek", + "max_input_tokens": 131072, + "max_output_tokens": 65536, + "max_tokens": 65536, + "mode": "chat", + "output_cost_per_token": 4.2e-07, + "source": "https://api-docs.deepseek.com/quick_start/pricing", + "supported_endpoints": [ + "/v1/chat/completions" + ], + "supports_function_calling": false, + "supports_native_streaming": true, + "supports_parallel_function_calling": false, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": false + }, "dashscope/qwen-coder": { "input_cost_per_token": 3e-07, "litellm_provider": "dashscope", @@ -9551,6 +9566,43 @@ } ] }, + "dashscope/qwen3-max-preview": { + "litellm_provider": "dashscope", + "max_input_tokens": 258048, + "max_output_tokens": 65536, + "max_tokens": 65536, + "mode": "chat", + "source": "https://www.alibabacloud.com/help/en/model-studio/models", + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "tiered_pricing": [ + { + "input_cost_per_token": 1.2e-06, + "output_cost_per_token": 6e-06, + "range": [ + 0, + 32000.0 + ] + }, + { + "input_cost_per_token": 2.4e-06, + "output_cost_per_token": 1.2e-05, + "range": [ + 32000.0, + 128000.0 + ] + }, + { + "input_cost_per_token": 3e-06, + "output_cost_per_token": 1.5e-05, + "range": [ + 128000.0, + 252000.0 + ] + } + ] + }, "dashscope/qwen3-max": { "litellm_provider": "dashscope", "max_input_tokens": 258048, @@ -9625,43 +9677,6 @@ } ] }, - "dashscope/qwen3-max-preview": { - "litellm_provider": "dashscope", - "max_input_tokens": 258048, - "max_output_tokens": 65536, - "max_tokens": 65536, - "mode": "chat", - "source": "https://www.alibabacloud.com/help/en/model-studio/models", - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true, - "tiered_pricing": [ - { - "input_cost_per_token": 1.2e-06, - "output_cost_per_token": 6e-06, - "range": [ - 0, - 32000.0 - ] - }, - { - "input_cost_per_token": 2.4e-06, - "output_cost_per_token": 1.2e-05, - "range": [ - 32000.0, - 128000.0 - ] - }, - { - "input_cost_per_token": 3e-06, - "output_cost_per_token": 1.5e-05, - "range": [ - 128000.0, - 252000.0 - ] - } - ] - }, "dashscope/qwen3-next-80b-a3b-instruct": { "input_cost_per_token": 1.5e-07, "litellm_provider": "dashscope", @@ -10300,40 +10315,6 @@ "mode": "completion", "output_cost_per_token": 2e-06 }, - "deep-research-pro-preview-12-2025": { - "input_cost_per_image": 0.0011, - "input_cost_per_token": 2e-06, - "input_cost_per_token_batches": 1e-06, - "litellm_provider": "vertex_ai-language-models", - "max_input_tokens": 65536, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "image_generation", - "output_cost_per_image": 0.134, - "output_cost_per_image_token": 0.00012, - "output_cost_per_token": 1.2e-05, - "output_cost_per_token_batches": 6e-06, - "source": "https://ai.google.dev/gemini-api/docs/pricing", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_function_calling": false, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_vision": true, - "supports_web_search": true - }, "deepgram/base": { "input_cost_per_second": 0.00020833, "litellm_provider": "deepgram", @@ -11566,88 +11547,6 @@ "supports_tool_choice": true, "supports_function_calling": true }, - "deepseek-chat": { - "cache_read_input_token_cost": 2.8e-08, - "input_cost_per_token": 2.8e-07, - "litellm_provider": "deepseek", - "max_input_tokens": 131072, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 4.2e-07, - "source": "https://api-docs.deepseek.com/quick_start/pricing", - "supported_endpoints": [ - "/v1/chat/completions" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "deepseek-reasoner": { - "cache_read_input_token_cost": 2.8e-08, - "input_cost_per_token": 2.8e-07, - "litellm_provider": "deepseek", - "max_input_tokens": 131072, - "max_output_tokens": 65536, - "max_tokens": 65536, - "mode": "chat", - "output_cost_per_token": 4.2e-07, - "source": "https://api-docs.deepseek.com/quick_start/pricing", - "supported_endpoints": [ - "/v1/chat/completions" - ], - "supports_function_calling": false, - "supports_native_streaming": true, - "supports_parallel_function_calling": false, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": false - }, - "deepseek-v3-2-251201": { - "input_cost_per_token": 0.0, - "litellm_provider": "volcengine", - "max_input_tokens": 98304, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "chat", - "output_cost_per_token": 0.0, - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, - "deepseek.v3-v1:0": { - "input_cost_per_token": 5.8e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 163840, - "max_output_tokens": 81920, - "max_tokens": 81920, - "mode": "chat", - "output_cost_per_token": 1.68e-06, - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, - "deepseek.v3.2": { - "input_cost_per_token": 6.2e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "max_tokens": 163840, - "mode": "chat", - "output_cost_per_token": 1.85e-06, - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, "deepseek/deepseek-chat": { "cache_creation_input_token_cost": 0.0, "cache_read_input_token_cost": 2.8e-08, @@ -11756,6 +11655,31 @@ "supports_reasoning": true, "supports_tool_choice": true }, + "deepseek.v3-v1:0": { + "input_cost_per_token": 5.8e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 163840, + "max_output_tokens": 81920, + "max_tokens": 81920, + "mode": "chat", + "output_cost_per_token": 1.68e-06, + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, + "deepseek.v3.2": { + "input_cost_per_token": 6.2e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "max_tokens": 163840, + "mode": "chat", + "output_cost_per_token": 1.85e-06, + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, "dolphin": { "input_cost_per_token": 5e-07, "litellm_provider": "nlp_cloud", @@ -11765,6 +11689,48 @@ "mode": "completion", "output_cost_per_token": 5e-07 }, + "deepseek-v3-2-251201": { + "input_cost_per_token": 0.0, + "litellm_provider": "volcengine", + "max_input_tokens": 98304, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "chat", + "output_cost_per_token": 0.0, + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, + "glm-4-7-251222": { + "input_cost_per_token": 0.0, + "litellm_provider": "volcengine", + "max_input_tokens": 204800, + "max_output_tokens": 131072, + "max_tokens": 131072, + "mode": "chat", + "output_cost_per_token": 0.0, + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, + "kimi-k2-thinking-251104": { + "input_cost_per_token": 0.0, + "litellm_provider": "volcengine", + "max_input_tokens": 229376, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "chat", + "output_cost_per_token": 0.0, + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, "doubao-embedding": { "input_cost_per_token": 0.0, "litellm_provider": "volcengine", @@ -11825,39 +11791,125 @@ "output_cost_per_token": 0.0, "output_vector_size": 2560 }, - "duckduckgo/search": { - "litellm_provider": "duckduckgo", + "exa_ai/search": { + "litellm_provider": "exa_ai", + "mode": "search", + "tiered_pricing": [ + { + "input_cost_per_query": 0.005, + "max_results_range": [ + 0, + 25 + ] + }, + { + "input_cost_per_query": 0.025, + "max_results_range": [ + 26, + 100 + ] + } + ] + }, + "firecrawl/search": { + "litellm_provider": "firecrawl", + "mode": "search", + "tiered_pricing": [ + { + "input_cost_per_query": 0.00166, + "max_results_range": [ + 1, + 10 + ] + }, + { + "input_cost_per_query": 0.00332, + "max_results_range": [ + 11, + 20 + ] + }, + { + "input_cost_per_query": 0.00498, + "max_results_range": [ + 21, + 30 + ] + }, + { + "input_cost_per_query": 0.00664, + "max_results_range": [ + 31, + 40 + ] + }, + { + "input_cost_per_query": 0.0083, + "max_results_range": [ + 41, + 50 + ] + }, + { + "input_cost_per_query": 0.00996, + "max_results_range": [ + 51, + 60 + ] + }, + { + "input_cost_per_query": 0.01162, + "max_results_range": [ + 61, + 70 + ] + }, + { + "input_cost_per_query": 0.01328, + "max_results_range": [ + 71, + 80 + ] + }, + { + "input_cost_per_query": 0.01494, + "max_results_range": [ + 81, + 90 + ] + }, + { + "input_cost_per_query": 0.0166, + "max_results_range": [ + 91, + 100 + ] + } + ], + "metadata": { + "notes": "Firecrawl search pricing: $83 for 100,000 credits, 2 credits per 10 results. Cost = ceiling(limit/10) * 2 * $0.00083" + } + }, + "perplexity/search": { + "input_cost_per_query": 0.005, + "litellm_provider": "perplexity", + "mode": "search" + }, + "searxng/search": { + "litellm_provider": "searxng", "mode": "search", "input_cost_per_query": 0.0, "metadata": { - "notes": "DuckDuckGo Instant Answer API is free and does not require an API key." + "notes": "SearXNG is an open-source metasearch engine. Free to use when self-hosted or using public instances." } }, - "elevenlabs/eleven_multilingual_v2": { - "input_cost_per_character": 0.00018, - "litellm_provider": "elevenlabs", + "serper/search": { + "input_cost_per_query": 0.001, + "litellm_provider": "serper", + "mode": "search", "metadata": { - "calculation": "$0.18/1000 characters (Scale plan pricing, 1 credit per character)", - "notes": "ElevenLabs Eleven Multilingual v2 - default TTS model with 29 languages support" - }, - "mode": "audio_speech", - "source": "https://elevenlabs.io/pricing", - "supported_endpoints": [ - "/v1/audio/speech" - ] - }, - "elevenlabs/eleven_v3": { - "input_cost_per_character": 0.00018, - "litellm_provider": "elevenlabs", - "metadata": { - "calculation": "$0.18/1000 characters (Scale plan pricing, 1 credit per character)", - "notes": "ElevenLabs Eleven v3 - most expressive TTS model with 70+ languages and audio tags support" - }, - "mode": "audio_speech", - "source": "https://elevenlabs.io/pricing", - "supported_endpoints": [ - "/v1/audio/speech" - ] + "notes": "Serper Google Search API. Pricing: $1.00/1k queries (Starter), $0.75/1k (Standard), $0.50/1k (Scale), $0.30/1k (Ultimate)." + } }, "elevenlabs/scribe_v1": { "input_cost_per_second": 6.11e-05, @@ -11889,6 +11941,32 @@ "/v1/audio/transcriptions" ] }, + "elevenlabs/eleven_v3": { + "input_cost_per_character": 0.00018, + "litellm_provider": "elevenlabs", + "metadata": { + "calculation": "$0.18/1000 characters (Scale plan pricing, 1 credit per character)", + "notes": "ElevenLabs Eleven v3 - most expressive TTS model with 70+ languages and audio tags support" + }, + "mode": "audio_speech", + "source": "https://elevenlabs.io/pricing", + "supported_endpoints": [ + "/v1/audio/speech" + ] + }, + "elevenlabs/eleven_multilingual_v2": { + "input_cost_per_character": 0.00018, + "litellm_provider": "elevenlabs", + "metadata": { + "calculation": "$0.18/1000 characters (Scale plan pricing, 1 credit per character)", + "notes": "ElevenLabs Eleven Multilingual v2 - default TTS model with 29 languages support" + }, + "mode": "audio_speech", + "source": "https://elevenlabs.io/pricing", + "supported_endpoints": [ + "/v1/audio/speech" + ] + }, "embed-english-light-v2.0": { "input_cost_per_token": 1e-07, "litellm_provider": "cohere", @@ -11927,15 +12005,6 @@ "supports_embedding_image_input": true, "supports_image_input": true }, - "embed-multilingual-light-v3.0": { - "input_cost_per_token": 0.0001, - "litellm_provider": "cohere", - "max_input_tokens": 1024, - "max_tokens": 1024, - "mode": "embedding", - "output_cost_per_token": 0.0, - "supports_embedding_image_input": true - }, "embed-multilingual-v2.0": { "input_cost_per_token": 1e-07, "litellm_provider": "cohere", @@ -11953,41 +12022,14 @@ "output_cost_per_token": 0.0, "supports_embedding_image_input": true }, - "eu.amazon.nova-2-lite-v1:0": { - "cache_read_input_token_cost": 8.25e-08, - "input_cost_per_token": 3.3e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 2.75e-06, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_video_input": true, - "supports_vision": true - }, - "eu.amazon.nova-2-pro-preview-20251202-v1:0": { - "cache_read_input_token_cost": 5.46875e-07, - "input_cost_per_token": 2.1875e-06, - "input_cost_per_image_token": 2.1875e-06, - "input_cost_per_audio_token": 2.1875e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.75e-05, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_video_input": true, - "supports_vision": true + "embed-multilingual-light-v3.0": { + "input_cost_per_token": 0.0001, + "litellm_provider": "cohere", + "max_input_tokens": 1024, + "max_tokens": 1024, + "mode": "embedding", + "output_cost_per_token": 0.0, + "supports_embedding_image_input": true }, "eu.amazon.nova-lite-v1:0": { "input_cost_per_token": 7.8e-08, @@ -12047,6 +12089,29 @@ "cache_read_input_token_cost": 2.5e-08, "cache_creation_input_token_cost": 3.125e-07 }, + "eu.anthropic.claude-haiku-4-5-20251001-v1:0": { + "cache_creation_input_token_cost": 1.375e-06, + "cache_read_input_token_cost": 1.1e-07, + "input_cost_per_token": 1.1e-06, + "deprecation_date": "2026-10-15", + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 5.5e-06, + "source": "https://aws.amazon.com/about-aws/whats-new/2025/10/claude-4-5-haiku-anthropic-amazon-bedrock", + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, "eu.anthropic.claude-3-5-sonnet-20240620-v1:0": { "input_cost_per_token": 3e-06, "litellm_provider": "bedrock", @@ -12149,29 +12214,6 @@ "cache_read_input_token_cost": 3e-07, "cache_creation_input_token_cost": 3.75e-06 }, - "eu.anthropic.claude-haiku-4-5-20251001-v1:0": { - "cache_creation_input_token_cost": 1.375e-06, - "cache_read_input_token_cost": 1.1e-07, - "input_cost_per_token": 1.1e-06, - "deprecation_date": "2026-10-15", - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 5.5e-06, - "source": "https://aws.amazon.com/about-aws/whats-new/2025/10/claude-4-5-haiku-anthropic-amazon-bedrock", - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, "eu.anthropic.claude-opus-4-1-20250805-v1:0": { "cache_creation_input_token_cost": 1.875e-05, "cache_read_input_token_cost": 1.5e-06, @@ -12224,62 +12266,6 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, - "eu.anthropic.claude-opus-4-5-20251101-v1:0": { - "cache_creation_input_token_cost": 6.25e-06, - "cache_read_input_token_cost": 5e-07, - "input_cost_per_token": 5e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 2.5e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 159 - }, - "eu.anthropic.claude-opus-4-6-v1": { - "cache_creation_input_token_cost": 6.875e-06, - "cache_creation_input_token_cost_above_200k_tokens": 1.375e-05, - "cache_read_input_token_cost": 5.5e-07, - "cache_read_input_token_cost_above_200k_tokens": 1.1e-06, - "input_cost_per_token": 5.5e-06, - "input_cost_per_token_above_200k_tokens": 1.1e-05, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 2.75e-05, - "output_cost_per_token_above_200k_tokens": 4.125e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": false, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, "eu.anthropic.claude-sonnet-4-20250514-v1:0": { "cache_creation_input_token_cost": 3.75e-06, "cache_read_input_token_cost": 3e-07, @@ -12340,48 +12326,6 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 346 }, - "eu.anthropic.claude-sonnet-4-6": { - "cache_creation_input_token_cost": 4.125e-06, - "cache_creation_input_token_cost_above_200k_tokens": 8.25e-06, - "cache_read_input_token_cost": 3.3e-07, - "cache_read_input_token_cost_above_200k_tokens": 6.6e-07, - "input_cost_per_token": 3.3e-06, - "input_cost_per_token_above_200k_tokens": 6.6e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.65e-05, - "output_cost_per_token_above_200k_tokens": 2.475e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, - "eu.deepseek.v3.2": { - "input_cost_per_token": 7.4e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "max_tokens": 163840, - "mode": "chat", - "output_cost_per_token": 2.22e-06, - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, "eu.meta.llama3-2-1b-instruct-v1:0": { "input_cost_per_token": 1.3e-07, "litellm_provider": "bedrock", @@ -12415,47 +12359,6 @@ "supports_function_calling": true, "supports_tool_choice": false }, - "eu.twelvelabs.marengo-embed-2-7-v1:0": { - "input_cost_per_token": 7e-05, - "input_cost_per_video_per_second": 0.0007, - "input_cost_per_audio_per_second": 0.00014, - "input_cost_per_image": 0.0001, - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "max_tokens": 77, - "mode": "embedding", - "output_cost_per_token": 0.0, - "output_vector_size": 1024, - "supports_embedding_image_input": true, - "supports_image_input": true - }, - "eu.twelvelabs.pegasus-1-2-v1:0": { - "input_cost_per_video_per_second": 0.00049, - "output_cost_per_token": 7.5e-06, - "litellm_provider": "bedrock", - "mode": "chat", - "supports_video_input": true - }, - "exa_ai/search": { - "litellm_provider": "exa_ai", - "mode": "search", - "tiered_pricing": [ - { - "input_cost_per_query": 0.005, - "max_results_range": [ - 0, - 25 - ] - }, - { - "input_cost_per_query": 0.025, - "max_results_range": [ - 26, - 100 - ] - } - ] - }, "fal_ai/bria/text-to-image/3.2": { "litellm_provider": "fal_ai", "mode": "image_generation", @@ -12464,22 +12367,6 @@ "/v1/images/generations" ] }, - "fal_ai/fal-ai/bytedance/dreamina/v3.1/text-to-image": { - "litellm_provider": "fal_ai", - "mode": "image_generation", - "output_cost_per_image": 0.03, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, - "fal_ai/fal-ai/bytedance/seedream/v3/text-to-image": { - "litellm_provider": "fal_ai", - "mode": "image_generation", - "output_cost_per_image": 0.03, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, "fal_ai/fal-ai/flux-pro/v1.1": { "litellm_provider": "fal_ai", "mode": "image_generation", @@ -12504,6 +12391,22 @@ "/v1/images/generations" ] }, + "fal_ai/fal-ai/bytedance/seedream/v3/text-to-image": { + "litellm_provider": "fal_ai", + "mode": "image_generation", + "output_cost_per_image": 0.03, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "fal_ai/fal-ai/bytedance/dreamina/v3.1/text-to-image": { + "litellm_provider": "fal_ai", + "mode": "image_generation", + "output_cost_per_image": 0.03, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, "fal_ai/fal-ai/ideogram/v3": { "litellm_provider": "fal_ai", "mode": "image_generation", @@ -12566,85 +12469,6 @@ "max_tokens": 4096, "mode": "chat" }, - "firecrawl/search": { - "litellm_provider": "firecrawl", - "mode": "search", - "tiered_pricing": [ - { - "input_cost_per_query": 0.00166, - "max_results_range": [ - 1, - 10 - ] - }, - { - "input_cost_per_query": 0.00332, - "max_results_range": [ - 11, - 20 - ] - }, - { - "input_cost_per_query": 0.00498, - "max_results_range": [ - 21, - 30 - ] - }, - { - "input_cost_per_query": 0.00664, - "max_results_range": [ - 31, - 40 - ] - }, - { - "input_cost_per_query": 0.0083, - "max_results_range": [ - 41, - 50 - ] - }, - { - "input_cost_per_query": 0.00996, - "max_results_range": [ - 51, - 60 - ] - }, - { - "input_cost_per_query": 0.01162, - "max_results_range": [ - 61, - 70 - ] - }, - { - "input_cost_per_query": 0.01328, - "max_results_range": [ - 71, - 80 - ] - }, - { - "input_cost_per_query": 0.01494, - "max_results_range": [ - 81, - 90 - ] - }, - { - "input_cost_per_query": 0.0166, - "max_results_range": [ - 91, - 100 - ] - } - ], - "metadata": { - "notes": "Firecrawl search pricing: $83 for 100,000 credits, 2 credits per 10 results. Cost = ceiling(limit/10) * 2 * $0.00083" - } - }, "fireworks-ai-4.1b-to-16b": { "input_cost_per_token": 2e-07, "litellm_provider": "fireworks_ai", @@ -12694,276 +12518,6 @@ "output_cost_per_token": 0.0, "source": "https://fireworks.ai/pricing" }, - "fireworks_ai/accounts/fireworks/models/": { - "max_tokens": 40960, - "max_input_tokens": 40960, - "max_output_tokens": 40960, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 0.0, - "litellm_provider": "fireworks_ai", - "mode": "embedding" - }, - "fireworks_ai/accounts/fireworks/models/SSD-1B": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1.3e-10, - "output_cost_per_token": 1.3e-10, - "litellm_provider": "fireworks_ai", - "mode": "image_generation" - }, - "fireworks_ai/accounts/fireworks/models/chronos-hermes-13b-v2": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-llama-13b": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-llama-13b-instruct": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-llama-13b-python": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-llama-34b": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-llama-34b-instruct": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-llama-34b-python": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-llama-70b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-llama-70b-instruct": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-llama-70b-python": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-llama-7b": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-llama-7b-instruct": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-llama-7b-python": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/code-qwen-1p5-7b": { - "max_tokens": 65536, - "max_input_tokens": 65536, - "max_output_tokens": 65536, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/codegemma-2b": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/codegemma-7b": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/cogito-671b-v2-p1": { - "max_tokens": 163840, - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "input_cost_per_token": 1.2e-06, - "output_cost_per_token": 1.2e-06, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/cogito-v1-preview-llama-3b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/cogito-v1-preview-llama-70b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/cogito-v1-preview-llama-8b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/cogito-v1-preview-qwen-14b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/cogito-v1-preview-qwen-32b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/dbrx-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1.2e-06, - "output_cost_per_token": 1.2e-06, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-coder-1b-base": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-coder-33b-instruct": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-coder-7b-base": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-coder-7b-base-v1p5": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-coder-7b-instruct-v1p5": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/deepseek-coder-v2-instruct": { "input_cost_per_token": 1.2e-06, "litellm_provider": "fireworks_ai", @@ -12977,33 +12531,6 @@ "supports_response_schema": true, "supports_tool_choice": false }, - "fireworks_ai/accounts/fireworks/models/deepseek-coder-v2-lite-base": { - "max_tokens": 163840, - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-coder-v2-lite-instruct": { - "max_tokens": 163840, - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-prover-v2": { - "max_tokens": 163840, - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "input_cost_per_token": 1.2e-06, - "output_cost_per_token": 1.2e-06, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/deepseek-r1": { "input_cost_per_token": 3e-06, "litellm_provider": "fireworks_ai", @@ -13028,15 +12555,6 @@ "supports_response_schema": true, "supports_tool_choice": false }, - "fireworks_ai/accounts/fireworks/models/deepseek-r1-0528-distill-qwen3-8b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/deepseek-r1-basic": { "input_cost_per_token": 5.5e-07, "litellm_provider": "fireworks_ai", @@ -13049,78 +12567,6 @@ "supports_response_schema": true, "supports_tool_choice": false }, - "fireworks_ai/accounts/fireworks/models/deepseek-r1-distill-llama-70b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-r1-distill-llama-8b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-r1-distill-qwen-14b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-r1-distill-qwen-1p5b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-r1-distill-qwen-32b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-r1-distill-qwen-7b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-v2-lite-chat": { - "max_tokens": 163840, - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/deepseek-v2p5": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1.2e-06, - "output_cost_per_token": 1.2e-06, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/deepseek-v3": { "input_cost_per_token": 9e-07, "litellm_provider": "fireworks_ai", @@ -13185,87 +12631,6 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "fireworks_ai/accounts/fireworks/models/devstral-small-2505": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/dobby-mini-unhinged-plus-llama-3-1-8b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/dobby-unhinged-llama-3-3-70b-new": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/dolphin-2-9-2-qwen2-72b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/dolphin-2p6-mixtral-8x7b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/ernie-4p5-21b-a3b-pt": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/ernie-4p5-300b-a47b-pt": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/fare-20b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/firefunction-v1": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/firefunction-v2": { "input_cost_per_token": 9e-07, "litellm_provider": "fireworks_ai", @@ -13279,150 +12644,6 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "fireworks_ai/accounts/fireworks/models/firellava-13b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/firesearch-ocr-v6": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/fireworks-asr-large": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 0.0, - "output_cost_per_token": 0.0, - "litellm_provider": "fireworks_ai", - "mode": "audio_transcription" - }, - "fireworks_ai/accounts/fireworks/models/fireworks-asr-v2": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 0.0, - "output_cost_per_token": 0.0, - "litellm_provider": "fireworks_ai", - "mode": "audio_transcription" - }, - "fireworks_ai/accounts/fireworks/models/flux-1-dev": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/flux-1-dev-controlnet-union": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1e-09, - "output_cost_per_token": 1e-09, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/flux-1-dev-fp8": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 5e-10, - "output_cost_per_token": 5e-10, - "litellm_provider": "fireworks_ai", - "mode": "image_generation" - }, - "fireworks_ai/accounts/fireworks/models/flux-1-schnell": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/flux-1-schnell-fp8": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 3.5e-10, - "output_cost_per_token": 3.5e-10, - "litellm_provider": "fireworks_ai", - "mode": "image_generation" - }, - "fireworks_ai/accounts/fireworks/models/flux-kontext-max": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 8e-08, - "output_cost_per_token": 8e-08, - "litellm_provider": "fireworks_ai", - "mode": "image_generation" - }, - "fireworks_ai/accounts/fireworks/models/flux-kontext-pro": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 4e-08, - "output_cost_per_token": 4e-08, - "litellm_provider": "fireworks_ai", - "mode": "image_generation" - }, - "fireworks_ai/accounts/fireworks/models/gemma-2b-it": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/gemma-3-27b-it": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/gemma-7b": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/gemma-7b-it": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/gemma2-9b-it": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/glm-4p5": { "input_cost_per_token": 5.5e-07, "litellm_provider": "fireworks_ai", @@ -13451,16 +12672,6 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "fireworks_ai/accounts/fireworks/models/glm-4p5v": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1.2e-06, - "output_cost_per_token": 1.2e-06, - "litellm_provider": "fireworks_ai", - "mode": "chat", - "supports_reasoning": true - }, "fireworks_ai/accounts/fireworks/models/glm-4p6": { "input_cost_per_token": 5.5e-07, "output_cost_per_token": 2.19e-06, @@ -13518,96 +12729,6 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "fireworks_ai/accounts/fireworks/models/gpt-oss-safeguard-120b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1.2e-06, - "output_cost_per_token": 1.2e-06, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/gpt-oss-safeguard-20b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/hermes-2-pro-mistral-7b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/internvl3-38b": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/internvl3-78b": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/internvl3-8b": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/japanese-stable-diffusion-xl": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1.3e-10, - "output_cost_per_token": 1.3e-10, - "litellm_provider": "fireworks_ai", - "mode": "image_generation" - }, - "fireworks_ai/accounts/fireworks/models/kat-coder": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/kat-dev-32b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/kat-dev-72b-exp": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/kimi-k2-instruct": { "input_cost_per_token": 6e-07, "litellm_provider": "fireworks_ai", @@ -13662,123 +12783,6 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "fireworks_ai/accounts/fireworks/models/llama-guard-2-8b": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-guard-3-1b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-guard-3-8b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-v2-13b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-v2-13b-chat": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-v2-70b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-v2-70b-chat": { - "max_tokens": 2048, - "max_input_tokens": 2048, - "max_output_tokens": 2048, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-v2-7b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-v2-7b-chat": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-v3-70b-instruct": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-v3-70b-instruct-hf": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-v3-8b": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-v3-8b-instruct-hf": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/llama-v3p1-405b-instruct": { "input_cost_per_token": 3e-06, "litellm_provider": "fireworks_ai", @@ -13792,33 +12796,6 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "fireworks_ai/accounts/fireworks/models/llama-v3p1-405b-instruct-long": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-v3p1-70b-instruct": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llama-v3p1-70b-instruct-1b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/llama-v3p1-8b-instruct": { "input_cost_per_token": 1e-07, "litellm_provider": "fireworks_ai", @@ -13832,15 +12809,6 @@ "supports_response_schema": true, "supports_tool_choice": false }, - "fireworks_ai/accounts/fireworks/models/llama-v3p1-nemotron-70b-instruct": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/llama-v3p2-11b-vision-instruct": { "input_cost_per_token": 2e-07, "litellm_provider": "fireworks_ai", @@ -13855,15 +12823,6 @@ "supports_tool_choice": false, "supports_vision": true }, - "fireworks_ai/accounts/fireworks/models/llama-v3p2-1b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/llama-v3p2-1b-instruct": { "input_cost_per_token": 1e-07, "litellm_provider": "fireworks_ai", @@ -13877,15 +12836,6 @@ "supports_response_schema": true, "supports_tool_choice": false }, - "fireworks_ai/accounts/fireworks/models/llama-v3p2-3b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/llama-v3p2-3b-instruct": { "input_cost_per_token": 1e-07, "litellm_provider": "fireworks_ai", @@ -13912,15 +12862,6 @@ "supports_tool_choice": false, "supports_vision": true }, - "fireworks_ai/accounts/fireworks/models/llama-v3p3-70b-instruct": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/llama4-maverick-instruct-basic": { "input_cost_per_token": 2.2e-07, "litellm_provider": "fireworks_ai", @@ -13945,42 +12886,6 @@ "supports_response_schema": true, "supports_tool_choice": false }, - "fireworks_ai/accounts/fireworks/models/llamaguard-7b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/llava-yi-34b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/minimax-m1-80k": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/minimax-m2": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 3e-07, - "output_cost_per_token": 1.2e-06, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/minimax-m2p1": { "cache_read_input_token_cost": 3e-08, "input_cost_per_token": 3e-07, @@ -13995,132 +12900,6 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "fireworks_ai/accounts/fireworks/models/ministral-3-14b-instruct-2512": { - "max_tokens": 256000, - "max_input_tokens": 256000, - "max_output_tokens": 256000, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/ministral-3-3b-instruct-2512": { - "max_tokens": 256000, - "max_input_tokens": 256000, - "max_output_tokens": 256000, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/ministral-3-8b-instruct-2512": { - "max_tokens": 256000, - "max_input_tokens": 256000, - "max_output_tokens": 256000, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mistral-7b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mistral-7b-instruct-4k": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mistral-7b-instruct-v0p2": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mistral-7b-instruct-v3": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mistral-7b-v0p2": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mistral-large-3-fp8": { - "max_tokens": 256000, - "max_input_tokens": 256000, - "max_output_tokens": 256000, - "input_cost_per_token": 1.2e-06, - "output_cost_per_token": 1.2e-06, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mistral-nemo-base-2407": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mistral-nemo-instruct-2407": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mistral-small-24b-instruct-2501": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mixtral-8x22b": { - "max_tokens": 65536, - "max_input_tokens": 65536, - "max_output_tokens": 65536, - "input_cost_per_token": 1.2e-06, - "output_cost_per_token": 1.2e-06, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mixtral-8x22b-instruct": { - "max_tokens": 65536, - "max_input_tokens": 65536, - "max_output_tokens": 65536, - "input_cost_per_token": 1.2e-06, - "output_cost_per_token": 1.2e-06, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/mixtral-8x22b-instruct-hf": { "input_cost_per_token": 1.2e-06, "litellm_provider": "fireworks_ai", @@ -14134,276 +12913,6 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "fireworks_ai/accounts/fireworks/models/mixtral-8x7b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mixtral-8x7b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mixtral-8x7b-instruct-hf": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/mythomax-l2-13b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/nemotron-nano-v2-12b-vl": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/nous-capybara-7b-v1p9": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/nous-hermes-2-mixtral-8x7b-dpo": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/nous-hermes-2-yi-34b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/nous-hermes-llama2-13b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/nous-hermes-llama2-70b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/nous-hermes-llama2-7b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/nvidia-nemotron-nano-12b-v2": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/nvidia-nemotron-nano-9b-v2": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/openchat-3p5-0106-7b": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/openhermes-2-mistral-7b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/openhermes-2p5-mistral-7b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/openorca-7b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/phi-2-3b": { - "max_tokens": 2048, - "max_input_tokens": 2048, - "max_output_tokens": 2048, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/phi-3-mini-128k-instruct": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/phi-3-vision-128k-instruct": { - "max_tokens": 32064, - "max_input_tokens": 32064, - "max_output_tokens": 32064, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/phind-code-llama-34b-python-v1": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/phind-code-llama-34b-v1": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/phind-code-llama-34b-v2": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/playground-v2-1024px-aesthetic": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1.3e-10, - "output_cost_per_token": 1.3e-10, - "litellm_provider": "fireworks_ai", - "mode": "image_generation" - }, - "fireworks_ai/accounts/fireworks/models/playground-v2-5-1024px-aesthetic": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1.3e-10, - "output_cost_per_token": 1.3e-10, - "litellm_provider": "fireworks_ai", - "mode": "image_generation" - }, - "fireworks_ai/accounts/fireworks/models/pythia-12b": { - "max_tokens": 2048, - "max_input_tokens": 2048, - "max_output_tokens": 2048, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen-qwq-32b-preview": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen-v2p5-14b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen-v2p5-7b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen1p5-72b-chat": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/qwen2-72b-instruct": { "input_cost_per_token": 9e-07, "litellm_provider": "fireworks_ai", @@ -14417,177 +12926,6 @@ "supports_response_schema": true, "supports_tool_choice": false }, - "fireworks_ai/accounts/fireworks/models/qwen2-7b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2-vl-2b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2-vl-72b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2-vl-7b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-0p5b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-14b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-1p5b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-32b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-32b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-72b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-72b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-7b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-0p5b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-0p5b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-14b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-14b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-1p5b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-1p5b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-32b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-32b-instruct": { "input_cost_per_token": 9e-07, "litellm_provider": "fireworks_ai", @@ -14601,558 +12939,6 @@ "supports_response_schema": true, "supports_tool_choice": false }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-32b-instruct-128k": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-32b-instruct-32k-rope": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-32b-instruct-64k": { - "max_tokens": 65536, - "max_input_tokens": 65536, - "max_output_tokens": 65536, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-3b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-3b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-7b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-7b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-math-72b-instruct": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-vl-32b-instruct": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-vl-3b-instruct": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-vl-72b-instruct": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen2p5-vl-7b-instruct": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-0p6b": { - "max_tokens": 40960, - "max_input_tokens": 40960, - "max_output_tokens": 40960, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-14b": { - "max_tokens": 40960, - "max_input_tokens": 40960, - "max_output_tokens": 40960, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-1p7b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-1p7b-fp8-draft": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-1p7b-fp8-draft-131072": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-1p7b-fp8-draft-40960": { - "max_tokens": 40960, - "max_input_tokens": 40960, - "max_output_tokens": 40960, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-235b-a22b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2.2e-07, - "output_cost_per_token": 8.8e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-235b-a22b-instruct-2507": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 2.2e-07, - "output_cost_per_token": 8.8e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-235b-a22b-thinking-2507": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 2.2e-07, - "output_cost_per_token": 8.8e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-30b-a3b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1.5e-07, - "output_cost_per_token": 6e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-30b-a3b-instruct-2507": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 5e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-30b-a3b-thinking-2507": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-32b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat", - "supports_reasoning": true - }, - "fireworks_ai/accounts/fireworks/models/qwen3-4b": { - "max_tokens": 40960, - "max_input_tokens": 40960, - "max_output_tokens": 40960, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-4b-instruct-2507": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-8b": { - "max_tokens": 40960, - "max_input_tokens": 40960, - "max_output_tokens": 40960, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat", - "supports_reasoning": true - }, - "fireworks_ai/accounts/fireworks/models/qwen3-coder-30b-a3b-instruct": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 1.5e-07, - "output_cost_per_token": 6e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-coder-480b-a35b-instruct": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 4.5e-07, - "output_cost_per_token": 1.8e-06, - "litellm_provider": "fireworks_ai", - "mode": "chat", - "supports_reasoning": true - }, - "fireworks_ai/accounts/fireworks/models/qwen3-coder-480b-instruct-bf16": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-embedding-0p6b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 0.0, - "output_cost_per_token": 0.0, - "litellm_provider": "fireworks_ai", - "mode": "embedding" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-embedding-4b": { - "max_tokens": 40960, - "max_input_tokens": 40960, - "max_output_tokens": 40960, - "input_cost_per_token": 0.0, - "output_cost_per_token": 0.0, - "litellm_provider": "fireworks_ai", - "mode": "embedding" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-next-80b-a3b-instruct": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-next-80b-a3b-thinking": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-reranker-0p6b": { - "max_tokens": 40960, - "max_input_tokens": 40960, - "max_output_tokens": 40960, - "input_cost_per_token": 0.0, - "output_cost_per_token": 0.0, - "litellm_provider": "fireworks_ai", - "mode": "rerank" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-reranker-4b": { - "max_tokens": 40960, - "max_input_tokens": 40960, - "max_output_tokens": 40960, - "input_cost_per_token": 0.0, - "output_cost_per_token": 0.0, - "litellm_provider": "fireworks_ai", - "mode": "rerank" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-reranker-8b": { - "max_tokens": 40960, - "max_input_tokens": 40960, - "max_output_tokens": 40960, - "input_cost_per_token": 0.0, - "output_cost_per_token": 0.0, - "litellm_provider": "fireworks_ai", - "mode": "rerank" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-vl-235b-a22b-instruct": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 2.2e-07, - "output_cost_per_token": 8.8e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-vl-235b-a22b-thinking": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 2.2e-07, - "output_cost_per_token": 8.8e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-vl-30b-a3b-instruct": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 1.5e-07, - "output_cost_per_token": 6e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-vl-30b-a3b-thinking": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 1.5e-07, - "output_cost_per_token": 6e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-vl-32b-instruct": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwen3-vl-8b-instruct": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/qwq-32b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/rolm-ocr": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/snorkel-mistral-7b-pairrm-dpo": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/stable-diffusion-xl-1024-v1-0": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1.3e-10, - "output_cost_per_token": 1.3e-10, - "litellm_provider": "fireworks_ai", - "mode": "image_generation" - }, - "fireworks_ai/accounts/fireworks/models/stablecode-3b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/starcoder-16b": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/starcoder-7b": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/starcoder2-15b": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/starcoder2-3b": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/starcoder2-7b": { - "max_tokens": 16384, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/toppy-m-7b": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/whisper-v3": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 0.0, - "output_cost_per_token": 0.0, - "litellm_provider": "fireworks_ai", - "mode": "audio_transcription" - }, - "fireworks_ai/accounts/fireworks/models/whisper-v3-turbo": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 0.0, - "output_cost_per_token": 0.0, - "litellm_provider": "fireworks_ai", - "mode": "audio_transcription" - }, - "fireworks_ai/accounts/fireworks/models/yi-34b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/yi-34b-200k-capybara": { - "max_tokens": 200000, - "max_input_tokens": 200000, - "max_output_tokens": 200000, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/yi-34b-chat": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 9e-07, - "output_cost_per_token": 9e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, - "fireworks_ai/accounts/fireworks/models/yi-6b": { - "max_tokens": 4096, - "max_input_tokens": 4096, - "max_output_tokens": 4096, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/accounts/fireworks/models/yi-large": { "input_cost_per_token": 3e-06, "litellm_provider": "fireworks_ai", @@ -15166,15 +12952,6 @@ "supports_response_schema": true, "supports_tool_choice": false }, - "fireworks_ai/accounts/fireworks/models/zephyr-7b-beta": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "fireworks_ai", - "mode": "chat" - }, "fireworks_ai/glm-4p7": { "cache_read_input_token_cost": 3e-07, "input_cost_per_token": 6e-07, @@ -15363,60 +13140,6 @@ "supports_system_messages": true, "supports_tool_choice": true }, - "ft:gpt-4.1-2025-04-14": { - "cache_read_input_token_cost": 7.5e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_batches": 1.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 1047576, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "chat", - "output_cost_per_token": 1.2e-05, - "output_cost_per_token_batches": 6e-06, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "ft:gpt-4.1-mini-2025-04-14": { - "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 8e-07, - "input_cost_per_token_batches": 4e-07, - "litellm_provider": "openai", - "max_input_tokens": 1047576, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "chat", - "output_cost_per_token": 3.2e-06, - "output_cost_per_token_batches": 1.6e-06, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "ft:gpt-4.1-nano-2025-04-14": { - "cache_read_input_token_cost": 5e-08, - "input_cost_per_token": 2e-07, - "input_cost_per_token_batches": 1e-07, - "litellm_provider": "openai", - "max_input_tokens": 1047576, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "chat", - "output_cost_per_token": 8e-07, - "output_cost_per_token_batches": 4e-07, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, "ft:gpt-4o-2024-08-06": { "cache_read_input_token_cost": 1.875e-06, "input_cost_per_token": 3.75e-06, @@ -15473,6 +13196,60 @@ "supports_system_messages": true, "supports_tool_choice": true }, + "ft:gpt-4.1-2025-04-14": { + "cache_read_input_token_cost": 7.5e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_batches": 1.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 1047576, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "chat", + "output_cost_per_token": 1.2e-05, + "output_cost_per_token_batches": 6e-06, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "ft:gpt-4.1-mini-2025-04-14": { + "cache_read_input_token_cost": 2e-07, + "input_cost_per_token": 8e-07, + "input_cost_per_token_batches": 4e-07, + "litellm_provider": "openai", + "max_input_tokens": 1047576, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "chat", + "output_cost_per_token": 3.2e-06, + "output_cost_per_token_batches": 1.6e-06, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "ft:gpt-4.1-nano-2025-04-14": { + "cache_read_input_token_cost": 5e-08, + "input_cost_per_token": 2e-07, + "input_cost_per_token_batches": 1e-07, + "litellm_provider": "openai", + "max_input_tokens": 1047576, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "chat", + "output_cost_per_token": 8e-07, + "output_cost_per_token_batches": 4e-07, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, "ft:o4-mini-2025-04-16": { "cache_read_input_token_cost": 1e-06, "input_cost_per_token": 4e-06, @@ -15569,27 +13346,6 @@ "supports_vision": true, "supports_web_search": true }, - "gemini-2.0-flash-exp-image-generation": { - "input_cost_per_token": 0.0, - "litellm_provider": "gemini", - "max_images_per_prompt": 3000, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "image_generation", - "output_cost_per_image": 0.039, - "output_cost_per_token": 0.0, - "source": "https://ai.google.dev/pricing", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_vision": true - }, "gemini-2.0-flash-lite": { "cache_read_input_token_cost": 1.875e-08, "deprecation_date": "2026-06-01", @@ -15662,31 +13418,6 @@ "supports_vision": true, "supports_web_search": true }, - "gemini-2.5-computer-use-preview-10-2025": { - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_above_200k_tokens": 2.5e-06, - "litellm_provider": "vertex_ai-language-models", - "max_images_per_prompt": 3000, - "max_input_tokens": 128000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_above_200k_tokens": 1.5e-05, - "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/computer-use", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_computer_use": true, - "supports_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true - }, "gemini-2.5-flash": { "cache_read_input_token_cost": 3e-08, "input_cost_per_audio_token": 1e-06, @@ -15781,24 +13512,91 @@ "supports_web_search": false, "tpm": 8000000 }, - "gemini-2.5-flash-lite": { - "cache_read_input_token_cost": 1e-08, - "input_cost_per_audio_token": 3e-07, - "input_cost_per_token": 1e-07, + "gemini-3-pro-image-preview": { + "input_cost_per_image": 0.0011, + "input_cost_per_token": 2e-06, + "input_cost_per_token_batches": 1e-06, + "litellm_provider": "vertex_ai-language-models", + "max_input_tokens": 65536, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "image_generation", + "output_cost_per_image": 0.134, + "output_cost_per_image_token": 0.00012, + "output_cost_per_token": 1.2e-05, + "output_cost_per_token_batches": 6e-06, + "source": "https://ai.google.dev/gemini-api/docs/pricing", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": false, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_vision": true, + "supports_web_search": true + }, + "gemini-3.1-flash-image-preview": { + "input_cost_per_image": 0.00056, + "input_cost_per_token": 5e-07, + "litellm_provider": "vertex_ai-language-models", + "max_input_tokens": 65536, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "image_generation", + "output_cost_per_image": 0.0672, + "output_cost_per_image_token": 6e-05, + "output_cost_per_token": 3e-06, + "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#gemini-models", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": false, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_vision": true, + "supports_web_search": true + }, + "gemini-3.1-flash-lite-preview": { + "cache_read_input_token_cost": 2.5e-08, + "cache_read_input_token_cost_per_audio_token": 5e-08, + "input_cost_per_audio_token": 5e-07, + "input_cost_per_token": 2.5e-07, "litellm_provider": "vertex_ai-language-models", "max_audio_length_hours": 8.4, "max_audio_per_prompt": 1, "max_images_per_prompt": 3000, "max_input_tokens": 1048576, - "max_output_tokens": 65535, + "max_output_tokens": 65536, "max_pdf_size_mb": 30, - "max_tokens": 65535, + "max_tokens": 65536, "max_video_length": 1, "max_videos_per_prompt": 10, "mode": "chat", - "output_cost_per_reasoning_token": 4e-07, - "output_cost_per_token": 4e-07, - "source": "https://ai.google.dev/gemini-api/docs/models#gemini-2.5-flash-preview", + "output_cost_per_reasoning_token": 1.5e-06, + "output_cost_per_token": 1.5e-06, + "source": "https://ai.google.dev/gemini-api/docs/models", "supported_endpoints": [ "/v1/chat/completions", "/v1/completions", @@ -15813,7 +13611,10 @@ "supported_output_modalities": [ "text" ], + "supports_audio_input": true, "supports_audio_output": false, + "supports_code_execution": true, + "supports_file_search": true, "supports_function_calling": true, "supports_parallel_function_calling": true, "supports_pdf_input": true, @@ -15823,13 +13624,48 @@ "supports_system_messages": true, "supports_tool_choice": true, "supports_url_context": true, + "supports_video_input": true, + "supports_vision": true, + "supports_web_search": true, + "supports_native_streaming": true + }, + "deep-research-pro-preview-12-2025": { + "input_cost_per_image": 0.0011, + "input_cost_per_token": 2e-06, + "input_cost_per_token_batches": 1e-06, + "litellm_provider": "vertex_ai-language-models", + "max_input_tokens": 65536, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "image_generation", + "output_cost_per_image": 0.134, + "output_cost_per_image_token": 0.00012, + "output_cost_per_token": 1.2e-05, + "output_cost_per_token_batches": 6e-06, + "source": "https://ai.google.dev/gemini-api/docs/pricing", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": false, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, "supports_vision": true, "supports_web_search": true }, - "gemini-2.5-flash-lite-preview-06-17": { - "deprecation_date": "2025-11-18", - "cache_read_input_token_cost": 2.5e-08, - "input_cost_per_audio_token": 5e-07, + "gemini-2.5-flash-lite": { + "cache_read_input_token_cost": 1e-08, + "input_cost_per_audio_token": 3e-07, "input_cost_per_token": 1e-07, "litellm_provider": "vertex_ai-language-models", "max_audio_length_hours": 8.4, @@ -15917,78 +13753,6 @@ "supports_vision": true, "supports_web_search": true }, - "gemini-2.5-flash-native-audio-latest": { - "input_cost_per_audio_token": 1e-06, - "input_cost_per_token": 3e-07, - "litellm_provider": "gemini", - "max_input_tokens": 1048576, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 2.5e-06, - "source": "https://ai.google.dev/pricing", - "supported_endpoints": [ - "/v1/realtime" - ], - "supported_modalities": [ - "text", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true - }, - "gemini-2.5-flash-native-audio-preview-09-2025": { - "input_cost_per_audio_token": 1e-06, - "input_cost_per_token": 3e-07, - "litellm_provider": "gemini", - "max_input_tokens": 1048576, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 2.5e-06, - "source": "https://ai.google.dev/pricing", - "supported_endpoints": [ - "/v1/realtime" - ], - "supported_modalities": [ - "text", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true - }, - "gemini-2.5-flash-native-audio-preview-12-2025": { - "input_cost_per_audio_token": 1e-06, - "input_cost_per_token": 3e-07, - "litellm_provider": "gemini", - "max_input_tokens": 1048576, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 2.5e-06, - "source": "https://ai.google.dev/pricing", - "supported_endpoints": [ - "/v1/realtime" - ], - "supported_modalities": [ - "text", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true - }, "gemini-2.5-flash-preview-09-2025": { "cache_read_input_token_cost": 7.5e-08, "input_cost_per_audio_token": 1e-06, @@ -16034,15 +13798,141 @@ "supports_vision": true, "supports_web_search": true }, - "gemini-2.5-flash-preview-tts": { + "gemini-live-2.5-flash-preview-native-audio-09-2025": { + "cache_read_input_token_cost": 7.5e-08, + "input_cost_per_audio_token": 3e-06, + "input_cost_per_token": 3e-07, + "litellm_provider": "vertex_ai-language-models", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "realtime", + "output_cost_per_audio_token": 1.2e-05, + "output_cost_per_token": 2e-06, + "source": "https://ai.google.dev/gemini-api/docs/pricing", + "supported_endpoints": [ + "/vertex_ai/live" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true + }, + "gemini/gemini-live-2.5-flash-preview-native-audio-09-2025": { + "cache_read_input_token_cost": 7.5e-08, + "input_cost_per_audio_token": 3e-06, "input_cost_per_token": 3e-07, "litellm_provider": "gemini", - "mode": "audio_speech", - "output_cost_per_token": 2.5e-06, - "source": "https://ai.google.dev/pricing", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "realtime", + "output_cost_per_audio_token": 1.2e-05, + "output_cost_per_token": 2e-06, + "rpm": 100000, + "source": "https://ai.google.dev/gemini-api/docs/pricing", "supported_endpoints": [ - "/v1/audio/speech" - ] + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true, + "tpm": 8000000 + }, + "gemini-2.5-flash-lite-preview-06-17": { + "deprecation_date": "2025-11-18", + "cache_read_input_token_cost": 2.5e-08, + "input_cost_per_audio_token": 5e-07, + "input_cost_per_token": 1e-07, + "litellm_provider": "vertex_ai-language-models", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_reasoning_token": 4e-07, + "output_cost_per_token": 4e-07, + "source": "https://ai.google.dev/gemini-api/docs/models#gemini-2.5-flash-preview", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true }, "gemini-2.5-pro": { "cache_read_input_token_cost": 1.25e-07, @@ -16089,127 +13979,6 @@ "supports_vision": true, "supports_web_search": true }, - "gemini-2.5-pro-preview-tts": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_above_200k_tokens": 2.5e-07, - "input_cost_per_audio_token": 7e-07, - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_above_200k_tokens": 2.5e-06, - "litellm_provider": "vertex_ai-language-models", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_above_200k_tokens": 1.5e-05, - "source": "https://ai.google.dev/gemini-api/docs/pricing#gemini-2.5-pro-preview", - "supported_modalities": [ - "text" - ], - "supported_output_modalities": [ - "audio" - ], - "supports_audio_output": false, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true - }, - "gemini-3-flash-preview": { - "cache_read_input_token_cost": 5e-08, - "input_cost_per_audio_token": 1e-06, - "input_cost_per_token": 5e-07, - "litellm_provider": "vertex_ai-language-models", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_reasoning_token": 3e-06, - "output_cost_per_token": 3e-06, - "source": "https://ai.google.dev/pricing/gemini-3", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_output": false, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_url_context": true, - "supports_vision": true, - "supports_web_search": true, - "supports_native_streaming": true, - "input_cost_per_token_priority": 9e-07, - "input_cost_per_audio_token_priority": 1.8e-06, - "output_cost_per_token_priority": 5.4e-06, - "cache_read_input_token_cost_priority": 9e-08, - "supports_service_tier": true - }, - "gemini-3-pro-image-preview": { - "input_cost_per_image": 0.0011, - "input_cost_per_token": 2e-06, - "input_cost_per_token_batches": 1e-06, - "litellm_provider": "vertex_ai-language-models", - "max_input_tokens": 65536, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "image_generation", - "output_cost_per_image": 0.134, - "output_cost_per_image_token": 0.00012, - "output_cost_per_token": 1.2e-05, - "output_cost_per_token_batches": 6e-06, - "source": "https://ai.google.dev/gemini-api/docs/pricing", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_function_calling": false, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_vision": true, - "supports_web_search": true - }, "gemini-3-pro-preview": { "deprecation_date": "2026-03-26", "cache_read_input_token_cost": 2e-07, @@ -16267,89 +14036,6 @@ "cache_read_input_token_cost_above_200k_tokens_priority": 7.2e-07, "supports_service_tier": true }, - "gemini-3.1-flash-image-preview": { - "input_cost_per_image": 0.00056, - "input_cost_per_token": 5e-07, - "litellm_provider": "vertex_ai-language-models", - "max_input_tokens": 65536, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "image_generation", - "output_cost_per_image": 0.0672, - "output_cost_per_image_token": 6e-05, - "output_cost_per_token": 3e-06, - "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#gemini-models", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_function_calling": false, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_vision": true, - "supports_web_search": true - }, - "gemini-3.1-flash-lite-preview": { - "cache_read_input_token_cost": 2.5e-08, - "cache_read_input_token_cost_per_audio_token": 5e-08, - "input_cost_per_audio_token": 5e-07, - "input_cost_per_token": 2.5e-07, - "litellm_provider": "vertex_ai-language-models", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65536, - "max_pdf_size_mb": 30, - "max_tokens": 65536, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_reasoning_token": 1.5e-06, - "output_cost_per_token": 1.5e-06, - "source": "https://ai.google.dev/gemini-api/docs/models", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_input": true, - "supports_audio_output": false, - "supports_code_execution": true, - "supports_file_search": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_url_context": true, - "supports_video_input": true, - "supports_vision": true, - "supports_web_search": true, - "supports_native_streaming": true - }, "gemini-3.1-pro-preview": { "cache_read_input_token_cost": 2e-07, "cache_read_input_token_cost_above_200k_tokens": 4e-07, @@ -16459,21 +14145,14 @@ "supports_url_context": true, "supports_native_streaming": true }, - "gemini-embedding-001": { - "input_cost_per_token": 1.5e-07, - "litellm_provider": "vertex_ai-embedding-models", - "max_input_tokens": 2048, - "max_tokens": 2048, - "mode": "embedding", - "output_cost_per_token": 0, - "output_vector_size": 3072, - "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models" - }, - "gemini-exp-1206": { - "cache_read_input_token_cost": 3e-08, - "input_cost_per_audio_token": 1e-06, - "input_cost_per_token": 3e-07, - "litellm_provider": "gemini", + "vertex_ai/gemini-3-pro-preview": { + "cache_read_input_token_cost": 2e-07, + "cache_read_input_token_cost_above_200k_tokens": 4e-07, + "cache_creation_input_token_cost_above_200k_tokens": 2.5e-07, + "input_cost_per_token": 2e-06, + "input_cost_per_token_above_200k_tokens": 4e-06, + "input_cost_per_token_batches": 1e-06, + "litellm_provider": "vertex_ai", "max_audio_length_hours": 8.4, "max_audio_per_prompt": 1, "max_images_per_prompt": 3000, @@ -16484,200 +14163,14 @@ "max_video_length": 1, "max_videos_per_prompt": 10, "mode": "chat", - "output_cost_per_reasoning_token": 2.5e-06, - "output_cost_per_token": 2.5e-06, - "rpm": 100000, - "source": "https://ai.google.dev/gemini-api/docs/models#gemini-2.5-flash-preview", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_output": false, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_url_context": true, - "supports_vision": true, - "supports_web_search": true, - "tpm": 8000000 - }, - "gemini-flash-latest": { - "cache_read_input_token_cost": 3e-08, - "input_cost_per_audio_token": 1e-06, - "input_cost_per_token": 3e-07, - "litellm_provider": "gemini", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_reasoning_token": 2.5e-06, - "output_cost_per_token": 2.5e-06, - "rpm": 100000, - "source": "https://ai.google.dev/gemini-api/docs/models#gemini-2.5-flash-preview", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_output": false, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_url_context": true, - "supports_vision": true, - "supports_web_search": true, - "tpm": 8000000 - }, - "gemini-flash-lite-latest": { - "cache_read_input_token_cost": 1e-08, - "input_cost_per_audio_token": 3e-07, - "input_cost_per_token": 1e-07, - "litellm_provider": "gemini", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_reasoning_token": 4e-07, - "output_cost_per_token": 4e-07, - "rpm": 15, - "source": "https://ai.google.dev/gemini-api/docs/models#gemini-2.5-flash-lite", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_output": false, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_url_context": true, - "supports_vision": true, - "supports_web_search": true, - "tpm": 250000 - }, - "gemini-live-2.5-flash-preview-native-audio-09-2025": { - "cache_read_input_token_cost": 7.5e-08, - "input_cost_per_audio_token": 3e-06, - "input_cost_per_token": 3e-07, - "litellm_provider": "vertex_ai-language-models", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_audio_token": 1.2e-05, - "output_cost_per_token": 2e-06, - "source": "https://ai.google.dev/gemini-api/docs/pricing", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_url_context": true, - "supports_vision": true, - "supports_web_search": true - }, - "gemini-pro-latest": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_above_200k_tokens": 2.5e-07, - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_above_200k_tokens": 2.5e-06, - "litellm_provider": "gemini", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_above_200k_tokens": 1.5e-05, - "rpm": 2000, + "output_cost_per_token": 1.2e-05, + "output_cost_per_token_above_200k_tokens": 1.8e-05, + "output_cost_per_token_batches": 6e-06, "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing", "supported_endpoints": [ "/v1/chat/completions", - "/v1/completions" + "/v1/completions", + "/v1/batch" ], "supported_modalities": [ "text", @@ -16699,7 +14192,215 @@ "supports_video_input": true, "supports_vision": true, "supports_web_search": true, - "tpm": 800000 + "supports_native_streaming": true, + "input_cost_per_token_priority": 3.6e-06, + "input_cost_per_token_above_200k_tokens_priority": 7.2e-06, + "output_cost_per_token_priority": 2.16e-05, + "output_cost_per_token_above_200k_tokens_priority": 3.24e-05, + "cache_read_input_token_cost_priority": 3.6e-07, + "cache_read_input_token_cost_above_200k_tokens_priority": 7.2e-07, + "supports_service_tier": true + }, + "vertex_ai/gemini-3-flash-preview": { + "cache_read_input_token_cost": 5e-08, + "input_cost_per_token": 5e-07, + "input_cost_per_audio_token": 1e-06, + "litellm_provider": "vertex_ai", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_token": 3e-06, + "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_input": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_video_input": true, + "supports_vision": true, + "supports_web_search": true, + "supports_native_streaming": true, + "input_cost_per_token_priority": 9e-07, + "input_cost_per_audio_token_priority": 1.8e-06, + "output_cost_per_token_priority": 5.4e-06, + "cache_read_input_token_cost_priority": 9e-08, + "supports_service_tier": true + }, + "vertex_ai/gemini-3.1-pro-preview": { + "cache_read_input_token_cost": 2e-07, + "cache_read_input_token_cost_above_200k_tokens": 4e-07, + "cache_creation_input_token_cost_above_200k_tokens": 2.5e-07, + "input_cost_per_token": 2e-06, + "input_cost_per_token_above_200k_tokens": 4e-06, + "input_cost_per_token_batches": 1e-06, + "litellm_provider": "vertex_ai", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65536, + "max_pdf_size_mb": 30, + "max_tokens": 65536, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_token": 1.2e-05, + "output_cost_per_token_above_200k_tokens": 1.8e-05, + "output_cost_per_token_batches": 6e-06, + "output_cost_per_image": 0.00012, + "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#gemini-models", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_input": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_video_input": true, + "supports_vision": true, + "supports_web_search": true, + "supports_url_context": true, + "supports_native_streaming": true, + "input_cost_per_token_priority": 3.6e-06, + "input_cost_per_token_above_200k_tokens_priority": 7.2e-06, + "output_cost_per_token_priority": 2.16e-05, + "output_cost_per_token_above_200k_tokens_priority": 3.24e-05, + "cache_read_input_token_cost_priority": 3.6e-07, + "cache_read_input_token_cost_above_200k_tokens_priority": 7.2e-07, + "supports_service_tier": true + }, + "vertex_ai/gemini-3.1-pro-preview-customtools": { + "cache_read_input_token_cost": 2e-07, + "cache_read_input_token_cost_above_200k_tokens": 4e-07, + "cache_creation_input_token_cost_above_200k_tokens": 2.5e-07, + "input_cost_per_token": 2e-06, + "input_cost_per_token_above_200k_tokens": 4e-06, + "input_cost_per_token_batches": 1e-06, + "litellm_provider": "vertex_ai", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65536, + "max_pdf_size_mb": 30, + "max_tokens": 65536, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_token": 1.2e-05, + "output_cost_per_token_above_200k_tokens": 1.8e-05, + "output_cost_per_token_batches": 6e-06, + "output_cost_per_image": 0.00012, + "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#gemini-models", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_input": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_video_input": true, + "supports_vision": true, + "supports_web_search": true, + "supports_url_context": true, + "supports_native_streaming": true, + "input_cost_per_token_priority": 3.6e-06, + "input_cost_per_token_above_200k_tokens_priority": 7.2e-06, + "output_cost_per_token_priority": 2.16e-05, + "output_cost_per_token_above_200k_tokens_priority": 3.24e-05, + "cache_read_input_token_cost_priority": 3.6e-07, + "cache_read_input_token_cost_above_200k_tokens_priority": 7.2e-07, + "supports_service_tier": true + }, + "gemini-2.5-pro-preview-tts": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_above_200k_tokens": 2.5e-07, + "input_cost_per_audio_token": 7e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_above_200k_tokens": 2.5e-06, + "litellm_provider": "vertex_ai-language-models", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_above_200k_tokens": 1.5e-05, + "source": "https://ai.google.dev/gemini-api/docs/pricing#gemini-2.5-pro-preview", + "supported_modalities": [ + "text" + ], + "supported_output_modalities": [ + "audio" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true }, "gemini-robotics-er-1.5-preview": { "cache_read_input_token_cost": 0, @@ -16737,41 +14438,130 @@ "supports_url_context": true, "supports_vision": true }, - "gemini/deep-research-pro-preview-12-2025": { - "input_cost_per_image": 0.0011, - "input_cost_per_token": 2e-06, - "input_cost_per_token_batches": 1e-06, + "gemini/gemini-robotics-er-1.5-preview": { + "cache_read_input_token_cost": 0, + "input_cost_per_token": 3e-07, + "input_cost_per_audio_token": 1e-06, "litellm_provider": "gemini", - "max_input_tokens": 65536, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "image_generation", - "output_cost_per_image": 0.134, - "output_cost_per_image_token": 0.00012, - "output_cost_per_token": 1.2e-05, - "rpm": 1000, - "tpm": 4000000, - "output_cost_per_token_batches": 6e-06, - "source": "https://ai.google.dev/gemini-api/docs/pricing", + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_tokens": 65535, + "mode": "chat", + "output_cost_per_token": 2.5e-06, + "output_cost_per_reasoning_token": 2.5e-06, + "source": "https://ai.google.dev/gemini-api/docs/models#gemini-robotics-er-1-5-preview", "supported_endpoints": [ "/v1/chat/completions", - "/v1/completions", - "/v1/batch" + "/v1/completions" ], + "supported_modalities": [ + "text", + "image", + "video", + "audio" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_prompt_caching": false, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true, + "tpm": 250000, + "rpm": 10 + }, + "gemini-2.5-computer-use-preview-10-2025": { + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_above_200k_tokens": 2.5e-06, + "litellm_provider": "vertex_ai-language-models", + "max_images_per_prompt": 3000, + "max_input_tokens": 128000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_above_200k_tokens": 1.5e-05, + "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/computer-use", "supported_modalities": [ "text", "image" ], "supported_output_modalities": [ - "text", - "image" + "text" ], - "supports_function_calling": false, - "supports_prompt_caching": true, - "supports_response_schema": true, + "supports_computer_use": true, + "supports_function_calling": true, "supports_system_messages": true, - "supports_vision": true, - "supports_web_search": true + "supports_tool_choice": true, + "supports_vision": true + }, + "gemini-embedding-001": { + "input_cost_per_token": 1.5e-07, + "litellm_provider": "vertex_ai-embedding-models", + "max_input_tokens": 2048, + "max_tokens": 2048, + "mode": "embedding", + "output_cost_per_token": 0, + "output_vector_size": 3072, + "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models" + }, + "gemini-embedding-2-preview": { + "input_cost_per_audio_per_second": 0.00016, + "input_cost_per_image": 0.00012, + "input_cost_per_token": 2e-07, + "input_cost_per_video_per_second": 0.0237, + "litellm_provider": "vertex_ai-embedding-models", + "max_input_tokens": 8192, + "max_tokens": 8192, + "mode": "embedding", + "output_cost_per_token": 0, + "output_vector_size": 3072, + "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing", + "uses_embed_content": true + }, + "vertex_ai/gemini-embedding-2-preview": { + "input_cost_per_token": 1.5e-07, + "litellm_provider": "vertex_ai", + "max_input_tokens": 8192, + "max_tokens": 8192, + "mode": "embedding", + "output_cost_per_token": 0, + "output_vector_size": 3072, + "source": "https://ai.google.dev/gemini-api/docs/embeddings#multimodal", + "supports_multimodal": true, + "uses_embed_content": true + }, + "gemini/gemini-embedding-001": { + "input_cost_per_token": 1.5e-07, + "litellm_provider": "gemini", + "max_input_tokens": 2048, + "max_tokens": 2048, + "mode": "embedding", + "output_cost_per_token": 0, + "output_vector_size": 3072, + "rpm": 10000, + "source": "https://ai.google.dev/gemini-api/docs/embeddings#model-versions", + "tpm": 10000000 + }, + "gemini/gemini-embedding-2-preview": { + "input_cost_per_token": 1.5e-07, + "litellm_provider": "gemini", + "max_input_tokens": 8192, + "max_tokens": 8192, + "mode": "embedding", + "output_cost_per_token": 0, + "output_vector_size": 3072, + "rpm": 10000, + "source": "https://ai.google.dev/gemini-api/docs/embeddings#multimodal", + "supports_multimodal": true, + "tpm": 10000000 }, "gemini/gemini-2.0-flash": { "cache_read_input_token_cost": 2.5e-08, @@ -16853,29 +14643,6 @@ "supports_web_search": true, "tpm": 10000000 }, - "gemini/gemini-2.0-flash-exp-image-generation": { - "input_cost_per_token": 0.0, - "litellm_provider": "gemini", - "max_images_per_prompt": 3000, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "image_generation", - "output_cost_per_image": 0.039, - "output_cost_per_token": 0.0, - "source": "https://ai.google.dev/pricing", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_vision": true, - "tpm": 250000, - "rpm": 10 - }, "gemini/gemini-2.0-flash-lite": { "cache_read_input_token_cost": 1.875e-08, "deprecation_date": "2026-06-01", @@ -16913,74 +14680,6 @@ "supports_web_search": true, "tpm": 4000000 }, - "gemini/gemini-2.0-flash-lite-001": { - "cache_read_input_token_cost": 1.875e-08, - "deprecation_date": "2026-06-01", - "input_cost_per_audio_token": 7.5e-08, - "input_cost_per_token": 7.5e-08, - "litellm_provider": "gemini", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 8192, - "max_pdf_size_mb": 50, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_token": 3e-07, - "rpm": 4000, - "source": "https://ai.google.dev/gemini-api/docs/pricing#gemini-2.0-flash-lite", - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_output": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "tpm": 4000000 - }, - "gemini/gemini-2.5-computer-use-preview-10-2025": { - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_above_200k_tokens": 2.5e-06, - "litellm_provider": "gemini", - "max_images_per_prompt": 3000, - "max_input_tokens": 128000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_above_200k_tokens": 1.5e-05, - "rpm": 2000, - "source": "https://ai.google.dev/gemini-api/docs/computer-use", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_computer_use": true, - "supports_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "tpm": 800000 - }, "gemini/gemini-2.5-flash": { "cache_read_input_token_cost": 3e-08, "input_cost_per_audio_token": 1e-06, @@ -17078,25 +14777,22 @@ "supports_web_search": true, "tpm": 8000000 }, - "gemini/gemini-2.5-flash-lite": { - "cache_read_input_token_cost": 1e-08, - "input_cost_per_audio_token": 3e-07, - "input_cost_per_token": 1e-07, + "gemini/gemini-3-pro-image-preview": { + "input_cost_per_image": 0.0011, + "input_cost_per_token": 2e-06, + "input_cost_per_token_batches": 1e-06, "litellm_provider": "gemini", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_reasoning_token": 4e-07, - "output_cost_per_token": 4e-07, - "rpm": 15, - "source": "https://ai.google.dev/gemini-api/docs/models#gemini-2.5-flash-lite", + "max_input_tokens": 65536, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "image_generation", + "output_cost_per_image": 0.134, + "output_cost_per_image_token": 0.00012, + "output_cost_per_token": 1.2e-05, + "rpm": 1000, + "tpm": 4000000, + "output_cost_per_token_batches": 6e-06, + "source": "https://ai.google.dev/gemini-api/docs/pricing", "supported_endpoints": [ "/v1/chat/completions", "/v1/completions", @@ -17104,31 +14800,94 @@ ], "supported_modalities": [ "text", - "image", - "audio", - "video" + "image" ], "supported_output_modalities": [ - "text" + "text", + "image" ], - "supports_audio_output": false, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, + "supports_function_calling": false, "supports_prompt_caching": true, - "supports_reasoning": true, "supports_response_schema": true, "supports_system_messages": true, - "supports_tool_choice": true, - "supports_url_context": true, "supports_vision": true, - "supports_web_search": true, - "tpm": 250000 + "supports_web_search": true }, - "gemini/gemini-2.5-flash-lite-preview-06-17": { - "deprecation_date": "2025-11-18", - "cache_read_input_token_cost": 2.5e-08, - "input_cost_per_audio_token": 5e-07, + "gemini/gemini-3.1-flash-image-preview": { + "input_cost_per_token": 2.5e-07, + "input_cost_per_token_batches": 1.25e-07, + "litellm_provider": "gemini", + "max_input_tokens": 65536, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "image_generation", + "output_cost_per_image": 0.045, + "output_cost_per_image_token": 6e-05, + "output_cost_per_image_token_batches": 3e-05, + "output_cost_per_token": 1.5e-06, + "output_cost_per_token_batches": 7.5e-07, + "rpm": 1000, + "tpm": 4000000, + "source": "https://ai.google.dev/gemini-api/docs/pricing#gemini-3.1-flash-image-preview", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": false, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_vision": true, + "supports_web_search": true + }, + "gemini/deep-research-pro-preview-12-2025": { + "input_cost_per_image": 0.0011, + "input_cost_per_token": 2e-06, + "input_cost_per_token_batches": 1e-06, + "litellm_provider": "gemini", + "max_input_tokens": 65536, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "image_generation", + "output_cost_per_image": 0.134, + "output_cost_per_image_token": 0.00012, + "output_cost_per_token": 1.2e-05, + "rpm": 1000, + "tpm": 4000000, + "output_cost_per_token_batches": 6e-06, + "source": "https://ai.google.dev/gemini-api/docs/pricing", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": false, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_vision": true, + "supports_web_search": true + }, + "gemini/gemini-2.5-flash-lite": { + "cache_read_input_token_cost": 1e-08, + "input_cost_per_audio_token": 3e-07, "input_cost_per_token": 1e-07, "litellm_provider": "gemini", "max_audio_length_hours": 8.4, @@ -17220,84 +14979,6 @@ "supports_web_search": true, "tpm": 250000 }, - "gemini/gemini-2.5-flash-native-audio-latest": { - "input_cost_per_audio_token": 1e-06, - "input_cost_per_token": 3e-07, - "litellm_provider": "gemini", - "max_input_tokens": 1048576, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 2.5e-06, - "source": "https://ai.google.dev/pricing", - "supported_endpoints": [ - "/v1/realtime" - ], - "supported_modalities": [ - "text", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true, - "tpm": 250000, - "rpm": 10 - }, - "gemini/gemini-2.5-flash-native-audio-preview-09-2025": { - "input_cost_per_audio_token": 1e-06, - "input_cost_per_token": 3e-07, - "litellm_provider": "gemini", - "max_input_tokens": 1048576, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 2.5e-06, - "source": "https://ai.google.dev/pricing", - "supported_endpoints": [ - "/v1/realtime" - ], - "supported_modalities": [ - "text", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true, - "tpm": 250000, - "rpm": 10 - }, - "gemini/gemini-2.5-flash-native-audio-preview-12-2025": { - "input_cost_per_audio_token": 1e-06, - "input_cost_per_token": 3e-07, - "litellm_provider": "gemini", - "max_input_tokens": 1048576, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 2.5e-06, - "source": "https://ai.google.dev/pricing", - "supported_endpoints": [ - "/v1/realtime" - ], - "supported_modalities": [ - "text", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true, - "tpm": 250000, - "rpm": 10 - }, "gemini/gemini-2.5-flash-preview-09-2025": { "cache_read_input_token_cost": 7.5e-08, "input_cost_per_audio_token": 1e-06, @@ -17345,6 +15026,148 @@ "supports_web_search": true, "tpm": 250000 }, + "gemini/gemini-flash-latest": { + "cache_read_input_token_cost": 7.5e-08, + "input_cost_per_audio_token": 1e-06, + "input_cost_per_token": 3e-07, + "litellm_provider": "gemini", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_reasoning_token": 2.5e-06, + "output_cost_per_token": 2.5e-06, + "rpm": 15, + "source": "https://developers.googleblog.com/en/continuing-to-bring-you-our-latest-models-with-an-improved-gemini-2-5-flash-and-flash-lite-release/", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true, + "tpm": 250000 + }, + "gemini/gemini-flash-lite-latest": { + "cache_read_input_token_cost": 2.5e-08, + "input_cost_per_audio_token": 3e-07, + "input_cost_per_token": 1e-07, + "litellm_provider": "gemini", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_reasoning_token": 4e-07, + "output_cost_per_token": 4e-07, + "rpm": 15, + "source": "https://developers.googleblog.com/en/continuing-to-bring-you-our-latest-models-with-an-improved-gemini-2-5-flash-and-flash-lite-release/", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true, + "tpm": 250000 + }, + "gemini/gemini-2.5-flash-lite-preview-06-17": { + "deprecation_date": "2025-11-18", + "cache_read_input_token_cost": 2.5e-08, + "input_cost_per_audio_token": 5e-07, + "input_cost_per_token": 1e-07, + "litellm_provider": "gemini", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_reasoning_token": 4e-07, + "output_cost_per_token": 4e-07, + "rpm": 15, + "source": "https://ai.google.dev/gemini-api/docs/models#gemini-2.5-flash-lite", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true, + "tpm": 250000 + }, "gemini/gemini-2.5-flash-preview-tts": { "input_cost_per_token": 3e-07, "litellm_provider": "gemini", @@ -17408,131 +15231,36 @@ "supports_web_search": true, "tpm": 800000 }, - "gemini/gemini-2.5-pro-preview-tts": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_above_200k_tokens": 2.5e-07, - "input_cost_per_audio_token": 7e-07, + "gemini/gemini-2.5-computer-use-preview-10-2025": { "input_cost_per_token": 1.25e-06, "input_cost_per_token_above_200k_tokens": 2.5e-06, "litellm_provider": "gemini", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, + "max_input_tokens": 128000, + "max_output_tokens": 64000, + "max_tokens": 64000, "mode": "chat", "output_cost_per_token": 1e-05, "output_cost_per_token_above_200k_tokens": 1.5e-05, - "rpm": 10000, - "source": "https://ai.google.dev/gemini-api/docs/pricing#gemini-2.5-pro-preview", - "supported_modalities": [ - "text" - ], - "supported_output_modalities": [ - "audio" - ], - "supports_audio_output": false, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "tpm": 10000000 - }, - "gemini/gemini-3-flash-preview": { - "cache_read_input_token_cost": 5e-08, - "input_cost_per_audio_token": 1e-06, - "input_cost_per_token": 5e-07, - "litellm_provider": "gemini", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_reasoning_token": 3e-06, - "output_cost_per_token": 3e-06, "rpm": 2000, - "source": "https://ai.google.dev/pricing/gemini-3", + "source": "https://ai.google.dev/gemini-api/docs/computer-use", "supported_endpoints": [ "/v1/chat/completions", - "/v1/completions", - "/v1/batch" + "/v1/completions" ], "supported_modalities": [ "text", - "image", - "audio", - "video" + "image" ], "supported_output_modalities": [ "text" ], - "supports_audio_output": false, + "supports_computer_use": true, "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, "supports_system_messages": true, "supports_tool_choice": true, - "supports_url_context": true, "supports_vision": true, - "supports_web_search": true, - "supports_native_streaming": true, - "tpm": 800000, - "input_cost_per_token_priority": 9e-07, - "input_cost_per_audio_token_priority": 1.8e-06, - "output_cost_per_token_priority": 5.4e-06, - "cache_read_input_token_cost_priority": 9e-08, - "supports_service_tier": true - }, - "gemini/gemini-3-pro-image-preview": { - "input_cost_per_image": 0.0011, - "input_cost_per_token": 2e-06, - "input_cost_per_token_batches": 1e-06, - "litellm_provider": "gemini", - "max_input_tokens": 65536, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "image_generation", - "output_cost_per_image": 0.134, - "output_cost_per_image_token": 0.00012, - "output_cost_per_token": 1.2e-05, - "rpm": 1000, - "tpm": 4000000, - "output_cost_per_token_batches": 6e-06, - "source": "https://ai.google.dev/gemini-api/docs/pricing", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_function_calling": false, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_vision": true, - "supports_web_search": true + "tpm": 800000 }, "gemini/gemini-3-pro-preview": { "deprecation_date": "2026-03-09", @@ -17591,42 +15319,6 @@ "cache_read_input_token_cost_above_200k_tokens_priority": 7.2e-07, "supports_service_tier": true }, - "gemini/gemini-3.1-flash-image-preview": { - "input_cost_per_token": 2.5e-07, - "input_cost_per_token_batches": 1.25e-07, - "litellm_provider": "gemini", - "max_input_tokens": 65536, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "image_generation", - "output_cost_per_image": 0.045, - "output_cost_per_image_token": 6e-05, - "output_cost_per_image_token_batches": 3e-05, - "output_cost_per_token": 1.5e-06, - "output_cost_per_token_batches": 7.5e-07, - "rpm": 1000, - "tpm": 4000000, - "source": "https://ai.google.dev/gemini-api/docs/pricing#gemini-3.1-flash-image-preview", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_function_calling": false, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_vision": true, - "supports_web_search": true - }, "gemini/gemini-3.1-flash-lite-preview": { "cache_read_input_token_cost": 2.5e-08, "cache_read_input_token_cost_per_audio_token": 5e-08, @@ -17680,6 +15372,59 @@ "supports_native_streaming": true, "tpm": 250000 }, + "gemini/gemini-3-flash-preview": { + "cache_read_input_token_cost": 5e-08, + "input_cost_per_audio_token": 1e-06, + "input_cost_per_token": 5e-07, + "litellm_provider": "gemini", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_reasoning_token": 3e-06, + "output_cost_per_token": 3e-06, + "rpm": 2000, + "source": "https://ai.google.dev/pricing/gemini-3", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true, + "supports_native_streaming": true, + "tpm": 800000, + "input_cost_per_token_priority": 9e-07, + "input_cost_per_audio_token_priority": 1.8e-06, + "output_cost_per_token_priority": 5.4e-06, + "cache_read_input_token_cost_priority": 9e-08, + "supports_service_tier": true + }, "gemini/gemini-3.1-pro-preview": { "cache_read_input_token_cost": 2e-07, "cache_read_input_token_cost_above_200k_tokens": 4e-07, @@ -17796,16 +15541,92 @@ "cache_read_input_token_cost_above_200k_tokens_priority": 7.2e-07, "supports_service_tier": true }, - "gemini/gemini-embedding-001": { - "input_cost_per_token": 1.5e-07, + "gemini-3-flash-preview": { + "cache_read_input_token_cost": 5e-08, + "input_cost_per_audio_token": 1e-06, + "input_cost_per_token": 5e-07, + "litellm_provider": "vertex_ai-language-models", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_reasoning_token": 3e-06, + "output_cost_per_token": 3e-06, + "source": "https://ai.google.dev/pricing/gemini-3", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true, + "supports_native_streaming": true, + "input_cost_per_token_priority": 9e-07, + "input_cost_per_audio_token_priority": 1.8e-06, + "output_cost_per_token_priority": 5.4e-06, + "cache_read_input_token_cost_priority": 9e-08, + "supports_service_tier": true + }, + "gemini/gemini-2.5-pro-preview-tts": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_above_200k_tokens": 2.5e-07, + "input_cost_per_audio_token": 7e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_above_200k_tokens": 2.5e-06, "litellm_provider": "gemini", - "max_input_tokens": 2048, - "max_tokens": 2048, - "mode": "embedding", - "output_cost_per_token": 0, - "output_vector_size": 3072, + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_above_200k_tokens": 1.5e-05, "rpm": 10000, - "source": "https://ai.google.dev/gemini-api/docs/embeddings#model-versions", + "source": "https://ai.google.dev/gemini-api/docs/pricing#gemini-2.5-pro-preview", + "supported_modalities": [ + "text" + ], + "supported_output_modalities": [ + "audio" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, "tpm": 10000000 }, "gemini/gemini-exp-1114": { @@ -17866,100 +15687,6 @@ "supports_vision": true, "tpm": 4000000 }, - "gemini/gemini-flash-latest": { - "cache_read_input_token_cost": 7.5e-08, - "input_cost_per_audio_token": 1e-06, - "input_cost_per_token": 3e-07, - "litellm_provider": "gemini", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_reasoning_token": 2.5e-06, - "output_cost_per_token": 2.5e-06, - "rpm": 15, - "source": "https://developers.googleblog.com/en/continuing-to-bring-you-our-latest-models-with-an-improved-gemini-2-5-flash-and-flash-lite-release/", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_output": false, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_url_context": true, - "supports_vision": true, - "supports_web_search": true, - "tpm": 250000 - }, - "gemini/gemini-flash-lite-latest": { - "cache_read_input_token_cost": 2.5e-08, - "input_cost_per_audio_token": 3e-07, - "input_cost_per_token": 1e-07, - "litellm_provider": "gemini", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_reasoning_token": 4e-07, - "output_cost_per_token": 4e-07, - "rpm": 15, - "source": "https://developers.googleblog.com/en/continuing-to-bring-you-our-latest-models-with-an-improved-gemini-2-5-flash-and-flash-lite-release/", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_output": false, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_url_context": true, - "supports_vision": true, - "supports_web_search": true, - "tpm": 250000 - }, "gemini/gemini-gemma-2-27b-it": { "input_cost_per_token": 3.5e-07, "litellm_provider": "gemini", @@ -17988,138 +15715,6 @@ "tpm": 250000, "rpm": 10 }, - "gemini/gemini-live-2.5-flash-preview-native-audio-09-2025": { - "cache_read_input_token_cost": 7.5e-08, - "input_cost_per_audio_token": 3e-06, - "input_cost_per_token": 3e-07, - "litellm_provider": "gemini", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_audio_token": 1.2e-05, - "output_cost_per_token": 2e-06, - "rpm": 100000, - "source": "https://ai.google.dev/gemini-api/docs/pricing", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_url_context": true, - "supports_vision": true, - "supports_web_search": true, - "tpm": 8000000 - }, - "gemini/gemini-pro-latest": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_above_200k_tokens": 2.5e-07, - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_above_200k_tokens": 2.5e-06, - "litellm_provider": "gemini", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_above_200k_tokens": 1.5e-05, - "rpm": 2000, - "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_input": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_video_input": true, - "supports_vision": true, - "supports_web_search": true, - "tpm": 800000 - }, - "gemini/gemini-robotics-er-1.5-preview": { - "cache_read_input_token_cost": 0, - "input_cost_per_token": 3e-07, - "input_cost_per_audio_token": 1e-06, - "litellm_provider": "gemini", - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_tokens": 65535, - "mode": "chat", - "output_cost_per_token": 2.5e-06, - "output_cost_per_reasoning_token": 2.5e-06, - "source": "https://ai.google.dev/gemini-api/docs/models#gemini-robotics-er-1-5-preview", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions" - ], - "supported_modalities": [ - "text", - "image", - "video", - "audio" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_output": false, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_prompt_caching": false, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_url_context": true, - "supports_vision": true, - "supports_web_search": true, - "tpm": 250000, - "rpm": 10 - }, "gemini/gemma-3-27b-it": { "input_cost_per_audio_per_second": 0, "input_cost_per_audio_per_second_above_128k_tokens": 0, @@ -18227,7 +15822,7 @@ "video" ] }, - "gemini/veo-3.1-fast-generate-001": { + "gemini/veo-3.1-fast-generate-preview": { "litellm_provider": "gemini", "max_input_tokens": 1024, "max_tokens": 1024, @@ -18241,7 +15836,21 @@ "video" ] }, - "gemini/veo-3.1-fast-generate-preview": { + "gemini/veo-3.1-generate-preview": { + "litellm_provider": "gemini", + "max_input_tokens": 1024, + "max_tokens": 1024, + "mode": "video_generation", + "output_cost_per_second": 0.4, + "source": "https://ai.google.dev/gemini-api/docs/video", + "supported_modalities": [ + "text" + ], + "supported_output_modalities": [ + "video" + ] + }, + "gemini/veo-3.1-fast-generate-001": { "litellm_provider": "gemini", "max_input_tokens": 1024, "max_tokens": 1024, @@ -18269,82 +15878,6 @@ "video" ] }, - "gemini/veo-3.1-generate-preview": { - "litellm_provider": "gemini", - "max_input_tokens": 1024, - "max_tokens": 1024, - "mode": "video_generation", - "output_cost_per_second": 0.4, - "source": "https://ai.google.dev/gemini-api/docs/video", - "supported_modalities": [ - "text" - ], - "supported_output_modalities": [ - "video" - ] - }, - "gigachat/Embeddings": { - "input_cost_per_token": 0.0, - "litellm_provider": "gigachat", - "max_input_tokens": 512, - "max_tokens": 512, - "mode": "embedding", - "output_cost_per_token": 0.0, - "output_vector_size": 1024 - }, - "gigachat/Embeddings-2": { - "input_cost_per_token": 0.0, - "litellm_provider": "gigachat", - "max_input_tokens": 512, - "max_tokens": 512, - "mode": "embedding", - "output_cost_per_token": 0.0, - "output_vector_size": 1024 - }, - "gigachat/EmbeddingsGigaR": { - "input_cost_per_token": 0.0, - "litellm_provider": "gigachat", - "max_input_tokens": 4096, - "max_tokens": 4096, - "mode": "embedding", - "output_cost_per_token": 0.0, - "output_vector_size": 2560 - }, - "gigachat/GigaChat-2-Lite": { - "input_cost_per_token": 0.0, - "litellm_provider": "gigachat", - "max_input_tokens": 128000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 0.0, - "supports_function_calling": true, - "supports_system_messages": true - }, - "gigachat/GigaChat-2-Max": { - "input_cost_per_token": 0.0, - "litellm_provider": "gigachat", - "max_input_tokens": 128000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 0.0, - "supports_function_calling": true, - "supports_system_messages": true, - "supports_vision": true - }, - "gigachat/GigaChat-2-Pro": { - "input_cost_per_token": 0.0, - "litellm_provider": "gigachat", - "max_input_tokens": 128000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 0.0, - "supports_function_calling": true, - "supports_system_messages": true, - "supports_vision": true - }, "github_copilot/claude-haiku-4.5": { "litellm_provider": "github_copilot", "max_input_tokens": 128000, @@ -18667,235 +16200,212 @@ "max_tokens": 8191, "mode": "embedding" }, - "glm-4-7-251222": { - "input_cost_per_token": 0.0, - "litellm_provider": "volcengine", - "max_input_tokens": 204800, - "max_output_tokens": 131072, - "max_tokens": 131072, - "mode": "chat", - "output_cost_per_token": 0.0, - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, - "global.amazon.nova-2-lite-v1:0": { - "cache_read_input_token_cost": 7.5e-08, - "input_cost_per_token": 3e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 2.5e-06, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_video_input": true, - "supports_vision": true - }, - "global.anthropic.claude-haiku-4-5-20251001-v1:0": { - "cache_creation_input_token_cost": 1.25e-06, - "cache_read_input_token_cost": 1e-07, - "input_cost_per_token": 1e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 5e-06, - "source": "https://aws.amazon.com/about-aws/whats-new/2025/10/claude-4-5-haiku-anthropic-amazon-bedrock", - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, - "global.anthropic.claude-opus-4-5-20251101-v1:0": { - "cache_creation_input_token_cost": 6.25e-06, - "cache_read_input_token_cost": 5e-07, - "input_cost_per_token": 5e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 2.5e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 159 - }, - "global.anthropic.claude-opus-4-6-v1": { - "cache_creation_input_token_cost": 6.25e-06, - "cache_creation_input_token_cost_above_200k_tokens": 1.25e-05, - "cache_read_input_token_cost": 5e-07, - "cache_read_input_token_cost_above_200k_tokens": 1e-06, - "input_cost_per_token": 5e-06, - "input_cost_per_token_above_200k_tokens": 1e-05, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, + "chatgpt/gpt-5.4": { + "litellm_provider": "chatgpt", + "max_input_tokens": 1050000, "max_output_tokens": 128000, "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 2.5e-05, - "output_cost_per_token_above_200k_tokens": 3.75e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": false, - "supports_computer_use": true, + "mode": "responses", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, + "supports_parallel_function_calling": true, "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, - "global.anthropic.claude-sonnet-4-20250514-v1:0": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 159 - }, - "global.anthropic.claude-sonnet-4-5-20250929-v1:0": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, - "global.anthropic.claude-sonnet-4-6": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost": 3e-07, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, - "gmi/MiniMaxAI/MiniMax-M2.1": { - "input_cost_per_token": 3e-07, - "litellm_provider": "gmi", - "max_input_tokens": 196608, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 1.2e-06 - }, - "gmi/Qwen/Qwen3-VL-235B-A22B-Instruct-FP8": { - "input_cost_per_token": 3e-07, - "litellm_provider": "gmi", - "max_input_tokens": 262144, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 1.4e-06, "supports_vision": true }, - "gmi/anthropic/claude-opus-4": { - "input_cost_per_token": 1.5e-05, - "litellm_provider": "gmi", - "max_input_tokens": 409600, - "max_output_tokens": 32000, - "max_tokens": 32000, - "mode": "chat", - "output_cost_per_token": 7.5e-05, + "chatgpt/gpt-5.4-pro": { + "litellm_provider": "chatgpt", + "max_input_tokens": 1050000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "supported_endpoints": [ + "/v1/responses" + ], "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, "supports_vision": true }, + "chatgpt/gpt-5.3-codex": { + "litellm_provider": "chatgpt", + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "supported_endpoints": [ + "/v1/responses" + ], + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, + "supports_vision": true + }, + "chatgpt/gpt-5.3-codex-spark": { + "litellm_provider": "chatgpt", + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "supported_endpoints": [ + "/v1/responses" + ], + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, + "supports_vision": true + }, + "chatgpt/gpt-5.3-instant": { + "litellm_provider": "chatgpt", + "max_input_tokens": 128000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "responses", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, + "supports_vision": true + }, + "chatgpt/gpt-5.3-chat-latest": { + "litellm_provider": "chatgpt", + "max_input_tokens": 128000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "responses", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, + "supports_vision": true + }, + "chatgpt/gpt-5.2-codex": { + "litellm_provider": "chatgpt", + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "supported_endpoints": [ + "/v1/responses" + ], + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, + "supports_vision": true + }, + "chatgpt/gpt-5.2": { + "litellm_provider": "chatgpt", + "max_input_tokens": 128000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "responses", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, + "supports_vision": true + }, + "chatgpt/gpt-5.1-codex-max": { + "litellm_provider": "chatgpt", + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "supported_endpoints": [ + "/v1/responses" + ], + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, + "supports_vision": true + }, + "chatgpt/gpt-5.1-codex-mini": { + "litellm_provider": "chatgpt", + "max_input_tokens": 128000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "responses", + "supported_endpoints": [ + "/v1/responses" + ], + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, + "supports_vision": true + }, + "gigachat/GigaChat-2-Lite": { + "input_cost_per_token": 0.0, + "litellm_provider": "gigachat", + "max_input_tokens": 128000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 0.0, + "supports_function_calling": true, + "supports_system_messages": true + }, + "gigachat/GigaChat-2-Max": { + "input_cost_per_token": 0.0, + "litellm_provider": "gigachat", + "max_input_tokens": 128000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 0.0, + "supports_function_calling": true, + "supports_system_messages": true, + "supports_vision": true + }, + "gigachat/GigaChat-2-Pro": { + "input_cost_per_token": 0.0, + "litellm_provider": "gigachat", + "max_input_tokens": 128000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 0.0, + "supports_function_calling": true, + "supports_system_messages": true, + "supports_vision": true + }, + "gigachat/Embeddings": { + "input_cost_per_token": 0.0, + "litellm_provider": "gigachat", + "max_input_tokens": 512, + "max_tokens": 512, + "mode": "embedding", + "output_cost_per_token": 0.0, + "output_vector_size": 1024 + }, + "gigachat/Embeddings-2": { + "input_cost_per_token": 0.0, + "litellm_provider": "gigachat", + "max_input_tokens": 512, + "max_tokens": 512, + "mode": "embedding", + "output_cost_per_token": 0.0, + "output_vector_size": 1024 + }, + "gigachat/EmbeddingsGigaR": { + "input_cost_per_token": 0.0, + "litellm_provider": "gigachat", + "max_input_tokens": 4096, + "max_tokens": 4096, + "mode": "embedding", + "output_cost_per_token": 0.0, + "output_vector_size": 2560 + }, "gmi/anthropic/claude-opus-4.5": { "input_cost_per_token": 5e-06, "litellm_provider": "gmi", @@ -18907,17 +16417,6 @@ "supports_function_calling": true, "supports_vision": true }, - "gmi/anthropic/claude-sonnet-4": { - "input_cost_per_token": 3e-06, - "litellm_provider": "gmi", - "max_input_tokens": 409600, - "max_output_tokens": 32000, - "max_tokens": 32000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "supports_function_calling": true, - "supports_vision": true - }, "gmi/anthropic/claude-sonnet-4.5": { "input_cost_per_token": 3e-06, "litellm_provider": "gmi", @@ -18929,56 +16428,57 @@ "supports_function_calling": true, "supports_vision": true }, - "gmi/deepseek-ai/DeepSeek-V3-0324": { - "input_cost_per_token": 2.8e-07, + "gmi/anthropic/claude-sonnet-4": { + "input_cost_per_token": 3e-06, "litellm_provider": "gmi", - "max_input_tokens": 163840, - "max_output_tokens": 16384, - "max_tokens": 16384, + "max_input_tokens": 409600, + "max_output_tokens": 32000, + "max_tokens": 32000, "mode": "chat", - "output_cost_per_token": 8.8e-07, - "supports_function_calling": true - }, - "gmi/deepseek-ai/DeepSeek-V3.2": { - "input_cost_per_token": 2.8e-07, - "litellm_provider": "gmi", - "max_input_tokens": 163840, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 4e-07, - "supports_function_calling": true - }, - "gmi/google/gemini-3-flash-preview": { - "input_cost_per_token": 5e-07, - "litellm_provider": "gmi", - "max_input_tokens": 1048576, - "max_output_tokens": 65536, - "max_tokens": 65536, - "mode": "chat", - "output_cost_per_token": 3e-06, + "output_cost_per_token": 1.5e-05, "supports_function_calling": true, "supports_vision": true }, - "gmi/google/gemini-3-pro-preview": { - "input_cost_per_token": 2e-06, + "gmi/anthropic/claude-opus-4": { + "input_cost_per_token": 1.5e-05, "litellm_provider": "gmi", - "max_input_tokens": 1048576, - "max_output_tokens": 65536, - "max_tokens": 65536, + "max_input_tokens": 409600, + "max_output_tokens": 32000, + "max_tokens": 32000, "mode": "chat", - "output_cost_per_token": 1.2e-05, + "output_cost_per_token": 7.5e-05, "supports_function_calling": true, "supports_vision": true }, - "gmi/moonshotai/Kimi-K2-Thinking": { - "input_cost_per_token": 8e-07, + "gmi/openai/gpt-5.2": { + "input_cost_per_token": 1.75e-06, "litellm_provider": "gmi", - "max_input_tokens": 262144, - "max_output_tokens": 16384, - "max_tokens": 16384, + "max_input_tokens": 409600, + "max_output_tokens": 32000, + "max_tokens": 32000, "mode": "chat", - "output_cost_per_token": 1.2e-06 + "output_cost_per_token": 1.4e-05, + "supports_function_calling": true + }, + "gmi/openai/gpt-5.1": { + "input_cost_per_token": 1.25e-06, + "litellm_provider": "gmi", + "max_input_tokens": 409600, + "max_output_tokens": 32000, + "max_tokens": 32000, + "mode": "chat", + "output_cost_per_token": 1e-05, + "supports_function_calling": true + }, + "gmi/openai/gpt-5": { + "input_cost_per_token": 1.25e-06, + "litellm_provider": "gmi", + "max_input_tokens": 409600, + "max_output_tokens": 32000, + "max_tokens": 32000, + "mode": "chat", + "output_cost_per_token": 1e-05, + "supports_function_calling": true }, "gmi/openai/gpt-4o": { "input_cost_per_token": 2.5e-06, @@ -19002,35 +16502,75 @@ "supports_function_calling": true, "supports_vision": true }, - "gmi/openai/gpt-5": { - "input_cost_per_token": 1.25e-06, + "gmi/deepseek-ai/DeepSeek-V3.2": { + "input_cost_per_token": 2.8e-07, "litellm_provider": "gmi", - "max_input_tokens": 409600, - "max_output_tokens": 32000, - "max_tokens": 32000, + "max_input_tokens": 163840, + "max_output_tokens": 16384, + "max_tokens": 16384, "mode": "chat", - "output_cost_per_token": 1e-05, + "output_cost_per_token": 4e-07, "supports_function_calling": true }, - "gmi/openai/gpt-5.1": { - "input_cost_per_token": 1.25e-06, + "gmi/deepseek-ai/DeepSeek-V3-0324": { + "input_cost_per_token": 2.8e-07, "litellm_provider": "gmi", - "max_input_tokens": 409600, - "max_output_tokens": 32000, - "max_tokens": 32000, + "max_input_tokens": 163840, + "max_output_tokens": 16384, + "max_tokens": 16384, "mode": "chat", - "output_cost_per_token": 1e-05, + "output_cost_per_token": 8.8e-07, "supports_function_calling": true }, - "gmi/openai/gpt-5.2": { - "input_cost_per_token": 1.75e-06, + "gmi/google/gemini-3-pro-preview": { + "input_cost_per_token": 2e-06, "litellm_provider": "gmi", - "max_input_tokens": 409600, - "max_output_tokens": 32000, - "max_tokens": 32000, + "max_input_tokens": 1048576, + "max_output_tokens": 65536, + "max_tokens": 65536, "mode": "chat", - "output_cost_per_token": 1.4e-05, - "supports_function_calling": true + "output_cost_per_token": 1.2e-05, + "supports_function_calling": true, + "supports_vision": true + }, + "gmi/google/gemini-3-flash-preview": { + "input_cost_per_token": 5e-07, + "litellm_provider": "gmi", + "max_input_tokens": 1048576, + "max_output_tokens": 65536, + "max_tokens": 65536, + "mode": "chat", + "output_cost_per_token": 3e-06, + "supports_function_calling": true, + "supports_vision": true + }, + "gmi/moonshotai/Kimi-K2-Thinking": { + "input_cost_per_token": 8e-07, + "litellm_provider": "gmi", + "max_input_tokens": 262144, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 1.2e-06 + }, + "gmi/MiniMaxAI/MiniMax-M2.1": { + "input_cost_per_token": 3e-07, + "litellm_provider": "gmi", + "max_input_tokens": 196608, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 1.2e-06 + }, + "gmi/Qwen/Qwen3-VL-235B-A22B-Instruct-FP8": { + "input_cost_per_token": 3e-07, + "litellm_provider": "gmi", + "max_input_tokens": 262144, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 1.4e-06, + "supports_vision": true }, "gmi/zai-org/GLM-4.7-FP8": { "input_cost_per_token": 4e-07, @@ -19079,6 +16619,105 @@ "litellm_provider": "google_pse", "mode": "search" }, + "global.anthropic.claude-sonnet-4-5-20250929-v1:0": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "global.anthropic.claude-sonnet-4-20250514-v1:0": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 159 + }, + "global.anthropic.claude-haiku-4-5-20251001-v1:0": { + "cache_creation_input_token_cost": 1.25e-06, + "cache_read_input_token_cost": 1e-07, + "input_cost_per_token": 1e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 5e-06, + "source": "https://aws.amazon.com/about-aws/whats-new/2025/10/claude-4-5-haiku-anthropic-amazon-bedrock", + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "global.amazon.nova-2-lite-v1:0": { + "cache_read_input_token_cost": 7.5e-08, + "input_cost_per_token": 3e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 1000000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 2.5e-06, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_video_input": true, + "supports_vision": true + }, "gpt-3.5-turbo": { "input_cost_per_token": 5e-07, "litellm_provider": "openai", @@ -19609,1551 +17248,6 @@ "supports_system_messages": true, "supports_tool_choice": true }, - "gpt-4o-mini": { - "cache_read_input_token_cost": 7.5e-08, - "cache_read_input_token_cost_priority": 1.25e-07, - "input_cost_per_token": 1.5e-07, - "input_cost_per_token_batches": 7.5e-08, - "input_cost_per_token_priority": 2.5e-07, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 6e-07, - "output_cost_per_token_batches": 3e-07, - "output_cost_per_token_priority": 1e-06, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true - }, - "gpt-4o-mini-2024-07-18": { - "cache_read_input_token_cost": 7.5e-08, - "input_cost_per_token": 1.5e-07, - "input_cost_per_token_batches": 7.5e-08, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 6e-07, - "output_cost_per_token_batches": 3e-07, - "search_context_cost_per_query": { - "search_context_size_high": 0.03, - "search_context_size_low": 0.025, - "search_context_size_medium": 0.0275 - }, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true - }, - "gpt-4o-mini-audio-preview": { - "input_cost_per_audio_token": 1e-05, - "input_cost_per_token": 1.5e-07, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_audio_token": 2e-05, - "output_cost_per_token": 6e-07, - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "gpt-4o-mini-audio-preview-2024-12-17": { - "input_cost_per_audio_token": 1e-05, - "input_cost_per_token": 1.5e-07, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_audio_token": 2e-05, - "output_cost_per_token": 6e-07, - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "gpt-4o-mini-realtime-preview": { - "cache_creation_input_audio_token_cost": 3e-07, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_audio_token": 1e-05, - "input_cost_per_token": 6e-07, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_audio_token": 2e-05, - "output_cost_per_token": 2.4e-06, - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "gpt-4o-mini-realtime-preview-2024-12-17": { - "cache_creation_input_audio_token_cost": 3e-07, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_audio_token": 1e-05, - "input_cost_per_token": 6e-07, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_audio_token": 2e-05, - "output_cost_per_token": 2.4e-06, - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "gpt-4o-mini-search-preview": { - "cache_read_input_token_cost": 7.5e-08, - "input_cost_per_token": 1.5e-07, - "input_cost_per_token_batches": 7.5e-08, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 6e-07, - "output_cost_per_token_batches": 3e-07, - "search_context_cost_per_query": { - "search_context_size_high": 0.03, - "search_context_size_low": 0.025, - "search_context_size_medium": 0.0275 - }, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true - }, - "gpt-4o-mini-search-preview-2025-03-11": { - "cache_read_input_token_cost": 7.5e-08, - "input_cost_per_token": 1.5e-07, - "input_cost_per_token_batches": 7.5e-08, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 6e-07, - "output_cost_per_token_batches": 3e-07, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true - }, - "gpt-4o-mini-transcribe": { - "input_cost_per_audio_token": 3e-06, - "input_cost_per_token": 1.25e-06, - "litellm_provider": "openai", - "max_input_tokens": 16000, - "max_output_tokens": 2000, - "mode": "audio_transcription", - "output_cost_per_token": 5e-06, - "supported_endpoints": [ - "/v1/audio/transcriptions" - ] - }, - "gpt-4o-mini-transcribe-2025-03-20": { - "input_cost_per_audio_token": 3e-06, - "input_cost_per_token": 1.25e-06, - "litellm_provider": "openai", - "max_input_tokens": 16000, - "max_output_tokens": 2000, - "mode": "audio_transcription", - "output_cost_per_token": 5e-06, - "supported_endpoints": [ - "/v1/audio/transcriptions" - ] - }, - "gpt-4o-mini-transcribe-2025-12-15": { - "input_cost_per_audio_token": 3e-06, - "input_cost_per_token": 1.25e-06, - "litellm_provider": "openai", - "max_input_tokens": 16000, - "max_output_tokens": 2000, - "mode": "audio_transcription", - "output_cost_per_token": 5e-06, - "supported_endpoints": [ - "/v1/audio/transcriptions" - ] - }, - "gpt-4o-mini-tts": { - "input_cost_per_token": 2.5e-06, - "litellm_provider": "openai", - "mode": "audio_speech", - "output_cost_per_audio_token": 1.2e-05, - "output_cost_per_second": 0.00025, - "output_cost_per_token": 1e-05, - "supported_endpoints": [ - "/v1/audio/speech" - ], - "supported_modalities": [ - "text", - "audio" - ], - "supported_output_modalities": [ - "audio" - ] - }, - "gpt-4o-mini-tts-2025-03-20": { - "input_cost_per_token": 2.5e-06, - "litellm_provider": "openai", - "mode": "audio_speech", - "output_cost_per_audio_token": 1.2e-05, - "output_cost_per_second": 0.00025, - "output_cost_per_token": 1e-05, - "supported_endpoints": [ - "/v1/audio/speech" - ], - "supported_modalities": [ - "text", - "audio" - ], - "supported_output_modalities": [ - "audio" - ] - }, - "gpt-4o-mini-tts-2025-12-15": { - "input_cost_per_token": 2.5e-06, - "litellm_provider": "openai", - "mode": "audio_speech", - "output_cost_per_audio_token": 1.2e-05, - "output_cost_per_second": 0.00025, - "output_cost_per_token": 1e-05, - "supported_endpoints": [ - "/v1/audio/speech" - ], - "supported_modalities": [ - "text", - "audio" - ], - "supported_output_modalities": [ - "audio" - ] - }, - "gpt-4o-realtime-preview": { - "cache_read_input_token_cost": 2.5e-06, - "input_cost_per_audio_token": 4e-05, - "input_cost_per_token": 5e-06, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_audio_token": 8e-05, - "output_cost_per_token": 2e-05, - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "gpt-4o-realtime-preview-2024-12-17": { - "cache_read_input_token_cost": 2.5e-06, - "input_cost_per_audio_token": 4e-05, - "input_cost_per_token": 5e-06, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_audio_token": 8e-05, - "output_cost_per_token": 2e-05, - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "gpt-4o-realtime-preview-2025-06-03": { - "cache_read_input_token_cost": 2.5e-06, - "input_cost_per_audio_token": 4e-05, - "input_cost_per_token": 5e-06, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_audio_token": 8e-05, - "output_cost_per_token": 2e-05, - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "gpt-4o-search-preview": { - "cache_read_input_token_cost": 1.25e-06, - "input_cost_per_token": 2.5e-06, - "input_cost_per_token_batches": 1.25e-06, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_batches": 5e-06, - "search_context_cost_per_query": { - "search_context_size_high": 0.05, - "search_context_size_low": 0.03, - "search_context_size_medium": 0.035 - }, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true - }, - "gpt-4o-search-preview-2025-03-11": { - "cache_read_input_token_cost": 1.25e-06, - "input_cost_per_token": 2.5e-06, - "input_cost_per_token_batches": 1.25e-06, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_batches": 5e-06, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true - }, - "gpt-4o-transcribe": { - "input_cost_per_audio_token": 6e-06, - "input_cost_per_token": 2.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 16000, - "max_output_tokens": 2000, - "mode": "audio_transcription", - "output_cost_per_token": 1e-05, - "supported_endpoints": [ - "/v1/audio/transcriptions" - ] - }, - "gpt-4o-transcribe-diarize": { - "input_cost_per_audio_token": 6e-06, - "input_cost_per_token": 2.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 16000, - "max_output_tokens": 2000, - "mode": "audio_transcription", - "output_cost_per_token": 1e-05, - "supported_endpoints": [ - "/v1/audio/transcriptions" - ] - }, - "gpt-5": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_flex": 6.25e-08, - "cache_read_input_token_cost_priority": 2.5e-07, - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_flex": 6.25e-07, - "input_cost_per_token_priority": 2.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_flex": 5e-06, - "output_cost_per_token_priority": 2e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5-2025-08-07": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_flex": 6.25e-08, - "cache_read_input_token_cost_priority": 2.5e-07, - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_flex": 6.25e-07, - "input_cost_per_token_priority": 2.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_flex": 5e-06, - "output_cost_per_token_priority": 2e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5-chat": { - "cache_read_input_token_cost": 1.25e-07, - "input_cost_per_token": 1.25e-06, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 1e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": false, - "supports_native_streaming": true, - "supports_parallel_function_calling": false, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": false, - "supports_vision": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5-chat-latest": { - "cache_read_input_token_cost": 1.25e-07, - "input_cost_per_token": 1.25e-06, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 1e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": false, - "supports_native_streaming": true, - "supports_parallel_function_calling": false, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": false, - "supports_vision": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5-codex": { - "cache_read_input_token_cost": 1.25e-07, - "input_cost_per_token": 1.25e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 1e-05, - "supported_endpoints": [ - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": false, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5-mini": { - "cache_read_input_token_cost": 2.5e-08, - "cache_read_input_token_cost_flex": 1.25e-08, - "cache_read_input_token_cost_priority": 4.5e-08, - "input_cost_per_token": 2.5e-07, - "input_cost_per_token_flex": 1.25e-07, - "input_cost_per_token_priority": 4.5e-07, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 2e-06, - "output_cost_per_token_flex": 1e-06, - "output_cost_per_token_priority": 3.6e-06, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5-mini-2025-08-07": { - "cache_read_input_token_cost": 2.5e-08, - "cache_read_input_token_cost_flex": 1.25e-08, - "cache_read_input_token_cost_priority": 4.5e-08, - "input_cost_per_token": 2.5e-07, - "input_cost_per_token_flex": 1.25e-07, - "input_cost_per_token_priority": 4.5e-07, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 2e-06, - "output_cost_per_token_flex": 1e-06, - "output_cost_per_token_priority": 3.6e-06, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5-nano": { - "cache_read_input_token_cost": 5e-09, - "cache_read_input_token_cost_flex": 2.5e-09, - "input_cost_per_token": 5e-08, - "input_cost_per_token_flex": 2.5e-08, - "input_cost_per_token_priority": 2.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 4e-07, - "output_cost_per_token_flex": 2e-07, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5-nano-2025-08-07": { - "cache_read_input_token_cost": 5e-09, - "cache_read_input_token_cost_flex": 2.5e-09, - "input_cost_per_token": 5e-08, - "input_cost_per_token_flex": 2.5e-08, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 4e-07, - "output_cost_per_token_flex": 2e-07, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5-pro": { - "input_cost_per_token": 1.5e-05, - "input_cost_per_token_batches": 7.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 272000, - "max_tokens": 272000, - "mode": "responses", - "output_cost_per_token": 0.00012, - "output_cost_per_token_batches": 6e-05, - "supported_endpoints": [ - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": false, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5-pro-2025-10-06": { - "input_cost_per_token": 1.5e-05, - "input_cost_per_token_batches": 7.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 272000, - "max_tokens": 272000, - "mode": "responses", - "output_cost_per_token": 0.00012, - "output_cost_per_token_batches": 6e-05, - "supported_endpoints": [ - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": false, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5-search-api": { - "cache_read_input_token_cost": 1.25e-07, - "input_cost_per_token": 1.25e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1e-05, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5-search-api-2025-10-14": { - "cache_read_input_token_cost": 1.25e-07, - "input_cost_per_token": 1.25e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1e-05, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5.1": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_priority": 2.5e-07, - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_priority": 2.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_priority": 2e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": true, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5.1-2025-11-13": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_priority": 2.5e-07, - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_priority": 2.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_priority": 2e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": true, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5.1-chat-latest": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_priority": 2.5e-07, - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_priority": 2.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 1e-05, - "output_cost_per_token_priority": 2e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_function_calling": false, - "supports_native_streaming": true, - "supports_parallel_function_calling": false, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": false, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": true, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5.1-codex": { - "cache_read_input_token_cost": 1.25e-07, - "cache_read_input_token_cost_priority": 2.5e-07, - "input_cost_per_token": 1.25e-06, - "input_cost_per_token_priority": 2.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 1e-05, - "output_cost_per_token_priority": 2e-05, - "supported_endpoints": [ - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": false, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5.1-codex-max": { - "cache_read_input_token_cost": 1.25e-07, - "input_cost_per_token": 1.25e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 1e-05, - "supported_endpoints": [ - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": false, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": true - }, - "gpt-5.1-codex-mini": { - "cache_read_input_token_cost": 2.5e-08, - "cache_read_input_token_cost_priority": 4.5e-08, - "input_cost_per_token": 2.5e-07, - "input_cost_per_token_priority": 4.5e-07, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 2e-06, - "output_cost_per_token_priority": 3.6e-06, - "supported_endpoints": [ - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": false, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5.2": { - "cache_read_input_token_cost": 1.75e-07, - "cache_read_input_token_cost_priority": 3.5e-07, - "input_cost_per_token": 1.75e-06, - "input_cost_per_token_priority": 3.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1.4e-05, - "output_cost_per_token_priority": 2.8e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": true, - "supports_xhigh_reasoning_effort": true - }, - "gpt-5.2-2025-12-11": { - "cache_read_input_token_cost": 1.75e-07, - "cache_read_input_token_cost_priority": 3.5e-07, - "input_cost_per_token": 1.75e-06, - "input_cost_per_token_priority": 3.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1.4e-05, - "output_cost_per_token_priority": 2.8e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text", - "image" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": true, - "supports_xhigh_reasoning_effort": true - }, - "gpt-5.2-chat-latest": { - "cache_read_input_token_cost": 1.75e-07, - "cache_read_input_token_cost_priority": 3.5e-07, - "input_cost_per_token": 1.75e-06, - "input_cost_per_token_priority": 3.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 1.4e-05, - "output_cost_per_token_priority": 2.8e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5.2-codex": { - "cache_read_input_token_cost": 1.75e-07, - "cache_read_input_token_cost_priority": 3.5e-07, - "input_cost_per_token": 1.75e-06, - "input_cost_per_token_priority": 3.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 1.4e-05, - "output_cost_per_token_priority": 2.8e-05, - "supported_endpoints": [ - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": false, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": true - }, - "gpt-5.2-pro": { - "input_cost_per_token": 2.1e-05, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 0.000168, - "supported_endpoints": [ - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": true - }, - "gpt-5.2-pro-2025-12-11": { - "input_cost_per_token": 2.1e-05, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 0.000168, - "supported_endpoints": [ - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": true - }, - "gpt-5.3-chat-latest": { - "cache_read_input_token_cost": 1.75e-07, - "cache_read_input_token_cost_priority": 3.5e-07, - "input_cost_per_token": 1.75e-06, - "input_cost_per_token_priority": 3.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 1.4e-05, - "output_cost_per_token_priority": 2.8e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5.3-codex": { - "cache_read_input_token_cost": 1.75e-07, - "cache_read_input_token_cost_priority": 3.5e-07, - "input_cost_per_token": 1.75e-06, - "input_cost_per_token_priority": 3.5e-06, - "litellm_provider": "openai", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 1.4e-05, - "output_cost_per_token_priority": 2.8e-05, - "supported_endpoints": [ - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": false, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": false - }, - "gpt-5.4": { - "cache_read_input_token_cost": 2.5e-07, - "cache_read_input_token_cost_above_272k_tokens": 5e-07, - "cache_read_input_token_cost_flex": 1.3e-07, - "cache_read_input_token_cost_priority": 5e-07, - "cache_read_input_token_cost_above_272k_tokens_priority": 1e-06, - "input_cost_per_token": 2.5e-06, - "input_cost_per_token_above_272k_tokens": 5e-06, - "input_cost_per_token_flex": 1.25e-06, - "input_cost_per_token_batches": 1.25e-06, - "input_cost_per_token_priority": 5e-06, - "input_cost_per_token_above_272k_tokens_priority": 1e-05, - "litellm_provider": "openai", - "max_input_tokens": 1050000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "output_cost_per_token_above_272k_tokens": 2.25e-05, - "output_cost_per_token_flex": 7.5e-06, - "output_cost_per_token_batches": 7.5e-06, - "output_cost_per_token_priority": 2.25e-05, - "output_cost_per_token_above_272k_tokens_priority": 3.375e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true, - "supports_none_reasoning_effort": true, - "supports_xhigh_reasoning_effort": true - }, - "gpt-5.4-2026-03-05": { - "cache_read_input_token_cost": 2.5e-07, - "cache_read_input_token_cost_above_272k_tokens": 5e-07, - "cache_read_input_token_cost_flex": 1.3e-07, - "cache_read_input_token_cost_priority": 5e-07, - "cache_read_input_token_cost_above_272k_tokens_priority": 1e-06, - "input_cost_per_token": 2.5e-06, - "input_cost_per_token_above_272k_tokens": 5e-06, - "input_cost_per_token_flex": 1.25e-06, - "input_cost_per_token_batches": 1.25e-06, - "input_cost_per_token_priority": 5e-06, - "input_cost_per_token_above_272k_tokens_priority": 1e-05, - "litellm_provider": "openai", - "max_input_tokens": 1050000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "output_cost_per_token_above_272k_tokens": 2.25e-05, - "output_cost_per_token_flex": 7.5e-06, - "output_cost_per_token_batches": 7.5e-06, - "output_cost_per_token_priority": 2.25e-05, - "output_cost_per_token_above_272k_tokens_priority": 3.375e-05, - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/batch", - "/v1/responses" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true - }, - "gpt-5.4-pro": { - "cache_read_input_token_cost": 3e-06, - "cache_read_input_token_cost_above_272k_tokens": 6e-06, - "cache_read_input_token_cost_priority": 6e-06, - "cache_read_input_token_cost_above_272k_tokens_priority": 1.2e-05, - "input_cost_per_token": 3e-05, - "input_cost_per_token_above_272k_tokens": 6e-05, - "input_cost_per_token_flex": 1.5e-05, - "input_cost_per_token_batches": 1.5e-05, - "input_cost_per_token_priority": 6e-05, - "input_cost_per_token_above_272k_tokens_priority": 0.00012, - "litellm_provider": "openai", - "max_input_tokens": 1050000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 0.00018, - "output_cost_per_token_above_272k_tokens": 0.00027, - "output_cost_per_token_flex": 9e-05, - "output_cost_per_token_batches": 9e-05, - "output_cost_per_token_priority": 0.00027, - "output_cost_per_token_above_272k_tokens_priority": 0.000405, - "supported_endpoints": [ - "/v1/responses", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": false, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": true - }, - "gpt-5.4-pro-2026-03-05": { - "cache_read_input_token_cost": 3e-06, - "cache_read_input_token_cost_above_272k_tokens": 6e-06, - "cache_read_input_token_cost_priority": 6e-06, - "cache_read_input_token_cost_above_272k_tokens_priority": 1.2e-05, - "input_cost_per_token": 3e-05, - "input_cost_per_token_above_272k_tokens": 6e-05, - "input_cost_per_token_flex": 1.5e-05, - "input_cost_per_token_batches": 1.5e-05, - "input_cost_per_token_priority": 6e-05, - "input_cost_per_token_above_272k_tokens_priority": 0.00012, - "litellm_provider": "openai", - "max_input_tokens": 1050000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "responses", - "output_cost_per_token": 0.00018, - "output_cost_per_token_above_272k_tokens": 0.00027, - "output_cost_per_token_flex": 9e-05, - "output_cost_per_token_batches": 9e-05, - "output_cost_per_token_priority": 0.00027, - "output_cost_per_token_above_272k_tokens_priority": 0.000405, - "supported_endpoints": [ - "/v1/responses", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_function_calling": true, - "supports_native_streaming": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": false, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_service_tier": true, - "supports_vision": true, - "supports_web_search": true, - "supports_none_reasoning_effort": false, - "supports_xhigh_reasoning_effort": true - }, "gpt-audio": { "input_cost_per_audio_token": 3.2e-05, "input_cost_per_token": 2.5e-06, @@ -21367,30 +17461,314 @@ "supports_tool_choice": true, "supports_vision": false }, - "gpt-image-1": { - "cache_read_input_image_token_cost": 2.5e-06, - "cache_read_input_token_cost": 1.25e-06, - "input_cost_per_image_token": 1e-05, - "input_cost_per_token": 5e-06, + "gpt-4o-mini": { + "cache_read_input_token_cost": 7.5e-08, + "cache_read_input_token_cost_priority": 1.25e-07, + "input_cost_per_token": 1.5e-07, + "input_cost_per_token_batches": 7.5e-08, + "input_cost_per_token_priority": 2.5e-07, "litellm_provider": "openai", - "mode": "image_generation", - "output_cost_per_image_token": 4e-05, + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 6e-07, + "output_cost_per_token_batches": 3e-07, + "output_cost_per_token_priority": 1e-06, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true + }, + "gpt-4o-mini-2024-07-18": { + "cache_read_input_token_cost": 7.5e-08, + "input_cost_per_token": 1.5e-07, + "input_cost_per_token_batches": 7.5e-08, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 6e-07, + "output_cost_per_token_batches": 3e-07, + "search_context_cost_per_query": { + "search_context_size_high": 0.03, + "search_context_size_low": 0.025, + "search_context_size_medium": 0.0275 + }, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true + }, + "gpt-4o-mini-audio-preview": { + "input_cost_per_audio_token": 1e-05, + "input_cost_per_token": 1.5e-07, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_audio_token": 2e-05, + "output_cost_per_token": 6e-07, + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "gpt-4o-mini-audio-preview-2024-12-17": { + "input_cost_per_audio_token": 1e-05, + "input_cost_per_token": 1.5e-07, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_audio_token": 2e-05, + "output_cost_per_token": 6e-07, + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "gpt-4o-mini-realtime-preview": { + "cache_creation_input_audio_token_cost": 3e-07, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_audio_token": 1e-05, + "input_cost_per_token": 6e-07, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_audio_token": 2e-05, + "output_cost_per_token": 2.4e-06, + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "gpt-4o-mini-realtime-preview-2024-12-17": { + "cache_creation_input_audio_token_cost": 3e-07, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_audio_token": 1e-05, + "input_cost_per_token": 6e-07, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_audio_token": 2e-05, + "output_cost_per_token": 2.4e-06, + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "gpt-4o-mini-search-preview": { + "cache_read_input_token_cost": 7.5e-08, + "input_cost_per_token": 1.5e-07, + "input_cost_per_token_batches": 7.5e-08, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 6e-07, + "output_cost_per_token_batches": 3e-07, + "search_context_cost_per_query": { + "search_context_size_high": 0.03, + "search_context_size_low": 0.025, + "search_context_size_medium": 0.0275 + }, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true + }, + "gpt-4o-mini-search-preview-2025-03-11": { + "cache_read_input_token_cost": 7.5e-08, + "input_cost_per_token": 1.5e-07, + "input_cost_per_token_batches": 7.5e-08, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 6e-07, + "output_cost_per_token_batches": 3e-07, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "gpt-4o-mini-transcribe": { + "input_cost_per_audio_token": 3e-06, + "input_cost_per_token": 1.25e-06, + "litellm_provider": "openai", + "max_input_tokens": 16000, + "max_output_tokens": 2000, + "mode": "audio_transcription", + "output_cost_per_token": 5e-06, "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" + "/v1/audio/transcriptions" ] }, - "gpt-image-1-mini": { - "cache_read_input_image_token_cost": 2.5e-07, - "cache_read_input_token_cost": 2e-07, - "input_cost_per_image_token": 2.5e-06, - "input_cost_per_token": 2e-06, + "gpt-4o-mini-tts": { + "input_cost_per_token": 2.5e-06, "litellm_provider": "openai", - "mode": "image_generation", - "output_cost_per_image_token": 8e-06, + "mode": "audio_speech", + "output_cost_per_audio_token": 1.2e-05, + "output_cost_per_second": 0.00025, + "output_cost_per_token": 1e-05, "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" + "/v1/audio/speech" + ], + "supported_modalities": [ + "text", + "audio" + ], + "supported_output_modalities": [ + "audio" + ] + }, + "gpt-4o-realtime-preview": { + "cache_read_input_token_cost": 2.5e-06, + "input_cost_per_audio_token": 4e-05, + "input_cost_per_token": 5e-06, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_audio_token": 8e-05, + "output_cost_per_token": 2e-05, + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "gpt-4o-realtime-preview-2024-12-17": { + "cache_read_input_token_cost": 2.5e-06, + "input_cost_per_audio_token": 4e-05, + "input_cost_per_token": 5e-06, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_audio_token": 8e-05, + "output_cost_per_token": 2e-05, + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "gpt-4o-realtime-preview-2025-06-03": { + "cache_read_input_token_cost": 2.5e-06, + "input_cost_per_audio_token": 4e-05, + "input_cost_per_token": 5e-06, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_audio_token": 8e-05, + "output_cost_per_token": 2e-05, + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "gpt-4o-search-preview": { + "cache_read_input_token_cost": 1.25e-06, + "input_cost_per_token": 2.5e-06, + "input_cost_per_token_batches": 1.25e-06, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_batches": 5e-06, + "search_context_cost_per_query": { + "search_context_size_high": 0.05, + "search_context_size_low": 0.03, + "search_context_size_medium": 0.035 + }, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true + }, + "gpt-4o-search-preview-2025-03-11": { + "cache_read_input_token_cost": 1.25e-06, + "input_cost_per_token": 2.5e-06, + "input_cost_per_token_batches": 1.25e-06, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_batches": 5e-06, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "gpt-4o-transcribe": { + "input_cost_per_audio_token": 6e-06, + "input_cost_per_token": 2.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 16000, + "max_output_tokens": 2000, + "mode": "audio_transcription", + "output_cost_per_token": 1e-05, + "supported_endpoints": [ + "/v1/audio/transcriptions" ] }, "gpt-image-1.5": { @@ -21423,6 +17801,1483 @@ "supports_vision": true, "supports_pdf_input": true }, + "low/1024-x-1024/gpt-image-1.5": { + "input_cost_per_image": 0.009, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "low/1024-x-1536/gpt-image-1.5": { + "input_cost_per_image": 0.013, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "low/1536-x-1024/gpt-image-1.5": { + "input_cost_per_image": 0.013, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "medium/1024-x-1024/gpt-image-1.5": { + "input_cost_per_image": 0.034, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "medium/1024-x-1536/gpt-image-1.5": { + "input_cost_per_image": 0.05, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "medium/1536-x-1024/gpt-image-1.5": { + "input_cost_per_image": 0.05, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "high/1024-x-1024/gpt-image-1.5": { + "input_cost_per_image": 0.133, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "high/1024-x-1536/gpt-image-1.5": { + "input_cost_per_image": 0.2, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "high/1536-x-1024/gpt-image-1.5": { + "input_cost_per_image": 0.2, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "standard/1024-x-1024/gpt-image-1.5": { + "input_cost_per_image": 0.009, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "standard/1024-x-1536/gpt-image-1.5": { + "input_cost_per_image": 0.013, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "standard/1536-x-1024/gpt-image-1.5": { + "input_cost_per_image": 0.013, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "1024-x-1024/gpt-image-1.5": { + "input_cost_per_image": 0.009, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "1024-x-1536/gpt-image-1.5": { + "input_cost_per_image": 0.013, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "1536-x-1024/gpt-image-1.5": { + "input_cost_per_image": 0.013, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "low/1024-x-1024/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.009, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "low/1024-x-1536/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.013, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "low/1536-x-1024/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.013, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "medium/1024-x-1024/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.034, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "medium/1024-x-1536/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.05, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "medium/1536-x-1024/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.05, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "high/1024-x-1024/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.133, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "high/1024-x-1536/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.2, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "high/1536-x-1024/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.2, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "standard/1024-x-1024/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.009, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "standard/1024-x-1536/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.013, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "standard/1536-x-1024/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.013, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "1024-x-1024/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.009, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "1024-x-1536/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.013, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "1536-x-1024/gpt-image-1.5-2025-12-16": { + "input_cost_per_image": 0.013, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ], + "supports_vision": true, + "supports_pdf_input": true + }, + "gpt-5": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_flex": 6.25e-08, + "cache_read_input_token_cost_priority": 2.5e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_flex": 6.25e-07, + "input_cost_per_token_priority": 2.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_flex": 5e-06, + "output_cost_per_token_priority": 2e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5.1": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_priority": 2.5e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_priority": 2.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_priority": 2e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": true, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5.1-2025-11-13": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_priority": 2.5e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_priority": 2.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_priority": 2e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": true, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5.1-chat-latest": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_priority": 2.5e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_priority": 2.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_priority": 2e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": false, + "supports_native_streaming": true, + "supports_parallel_function_calling": false, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": false, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": true, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5.2": { + "cache_read_input_token_cost": 1.75e-07, + "cache_read_input_token_cost_priority": 3.5e-07, + "input_cost_per_token": 1.75e-06, + "input_cost_per_token_priority": 3.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1.4e-05, + "output_cost_per_token_priority": 2.8e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": true, + "supports_xhigh_reasoning_effort": true + }, + "gpt-5.2-2025-12-11": { + "cache_read_input_token_cost": 1.75e-07, + "cache_read_input_token_cost_priority": 3.5e-07, + "input_cost_per_token": 1.75e-06, + "input_cost_per_token_priority": 3.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1.4e-05, + "output_cost_per_token_priority": 2.8e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": true, + "supports_xhigh_reasoning_effort": true + }, + "gpt-5.2-chat-latest": { + "cache_read_input_token_cost": 1.75e-07, + "cache_read_input_token_cost_priority": 3.5e-07, + "input_cost_per_token": 1.75e-06, + "input_cost_per_token_priority": 3.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 1.4e-05, + "output_cost_per_token_priority": 2.8e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5.3-chat-latest": { + "cache_read_input_token_cost": 1.75e-07, + "cache_read_input_token_cost_priority": 3.5e-07, + "input_cost_per_token": 1.75e-06, + "input_cost_per_token_priority": 3.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 1.4e-05, + "output_cost_per_token_priority": 2.8e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5.2-pro": { + "input_cost_per_token": 2.1e-05, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 0.000168, + "supported_endpoints": [ + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": true + }, + "gpt-5.2-pro-2025-12-11": { + "input_cost_per_token": 2.1e-05, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 0.000168, + "supported_endpoints": [ + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": true + }, + "gpt-5.4": { + "cache_read_input_token_cost": 2.5e-07, + "cache_read_input_token_cost_above_272k_tokens": 5e-07, + "cache_read_input_token_cost_flex": 1.3e-07, + "cache_read_input_token_cost_priority": 5e-07, + "cache_read_input_token_cost_above_272k_tokens_priority": 1e-06, + "input_cost_per_token": 2.5e-06, + "input_cost_per_token_above_272k_tokens": 5e-06, + "input_cost_per_token_flex": 1.25e-06, + "input_cost_per_token_batches": 1.25e-06, + "input_cost_per_token_priority": 5e-06, + "input_cost_per_token_above_272k_tokens_priority": 1e-05, + "litellm_provider": "openai", + "max_input_tokens": 1050000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "output_cost_per_token_above_272k_tokens": 2.25e-05, + "output_cost_per_token_flex": 7.5e-06, + "output_cost_per_token_batches": 7.5e-06, + "output_cost_per_token_priority": 2.25e-05, + "output_cost_per_token_above_272k_tokens_priority": 3.375e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true, + "supports_none_reasoning_effort": true, + "supports_xhigh_reasoning_effort": true + }, + "gpt-5.4-2026-03-05": { + "cache_read_input_token_cost": 2.5e-07, + "cache_read_input_token_cost_above_272k_tokens": 5e-07, + "cache_read_input_token_cost_flex": 1.3e-07, + "cache_read_input_token_cost_priority": 5e-07, + "cache_read_input_token_cost_above_272k_tokens_priority": 1e-06, + "input_cost_per_token": 2.5e-06, + "input_cost_per_token_above_272k_tokens": 5e-06, + "input_cost_per_token_flex": 1.25e-06, + "input_cost_per_token_batches": 1.25e-06, + "input_cost_per_token_priority": 5e-06, + "input_cost_per_token_above_272k_tokens_priority": 1e-05, + "litellm_provider": "openai", + "max_input_tokens": 1050000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "output_cost_per_token_above_272k_tokens": 2.25e-05, + "output_cost_per_token_flex": 7.5e-06, + "output_cost_per_token_batches": 7.5e-06, + "output_cost_per_token_priority": 2.25e-05, + "output_cost_per_token_above_272k_tokens_priority": 3.375e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true + }, + "gpt-5.4-pro": { + "cache_read_input_token_cost": 3e-06, + "cache_read_input_token_cost_above_272k_tokens": 6e-06, + "cache_read_input_token_cost_priority": 6e-06, + "cache_read_input_token_cost_above_272k_tokens_priority": 1.2e-05, + "input_cost_per_token": 3e-05, + "input_cost_per_token_above_272k_tokens": 6e-05, + "input_cost_per_token_flex": 1.5e-05, + "input_cost_per_token_batches": 1.5e-05, + "input_cost_per_token_priority": 6e-05, + "input_cost_per_token_above_272k_tokens_priority": 0.00012, + "litellm_provider": "openai", + "max_input_tokens": 1050000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 0.00018, + "output_cost_per_token_above_272k_tokens": 0.00027, + "output_cost_per_token_flex": 9e-05, + "output_cost_per_token_batches": 9e-05, + "output_cost_per_token_priority": 0.00027, + "output_cost_per_token_above_272k_tokens_priority": 0.000405, + "supported_endpoints": [ + "/v1/responses", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": false, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": true + }, + "gpt-5.4-pro-2026-03-05": { + "cache_read_input_token_cost": 3e-06, + "cache_read_input_token_cost_above_272k_tokens": 6e-06, + "cache_read_input_token_cost_priority": 6e-06, + "cache_read_input_token_cost_above_272k_tokens_priority": 1.2e-05, + "input_cost_per_token": 3e-05, + "input_cost_per_token_above_272k_tokens": 6e-05, + "input_cost_per_token_flex": 1.5e-05, + "input_cost_per_token_batches": 1.5e-05, + "input_cost_per_token_priority": 6e-05, + "input_cost_per_token_above_272k_tokens_priority": 0.00012, + "litellm_provider": "openai", + "max_input_tokens": 1050000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 0.00018, + "output_cost_per_token_above_272k_tokens": 0.00027, + "output_cost_per_token_flex": 9e-05, + "output_cost_per_token_batches": 9e-05, + "output_cost_per_token_priority": 0.00027, + "output_cost_per_token_above_272k_tokens_priority": 0.000405, + "supported_endpoints": [ + "/v1/responses", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": false, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": true + }, + "gpt-5-pro": { + "input_cost_per_token": 1.5e-05, + "input_cost_per_token_batches": 7.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 272000, + "max_tokens": 272000, + "mode": "responses", + "output_cost_per_token": 0.00012, + "output_cost_per_token_batches": 6e-05, + "supported_endpoints": [ + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": false, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5-pro-2025-10-06": { + "input_cost_per_token": 1.5e-05, + "input_cost_per_token_batches": 7.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 272000, + "max_tokens": 272000, + "mode": "responses", + "output_cost_per_token": 0.00012, + "output_cost_per_token_batches": 6e-05, + "supported_endpoints": [ + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": false, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5-2025-08-07": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_flex": 6.25e-08, + "cache_read_input_token_cost_priority": 2.5e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_flex": 6.25e-07, + "input_cost_per_token_priority": 2.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_flex": 5e-06, + "output_cost_per_token_priority": 2e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5-chat": { + "cache_read_input_token_cost": 1.25e-07, + "input_cost_per_token": 1.25e-06, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 1e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": false, + "supports_native_streaming": true, + "supports_parallel_function_calling": false, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": false, + "supports_vision": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5-chat-latest": { + "cache_read_input_token_cost": 1.25e-07, + "input_cost_per_token": 1.25e-06, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 1e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": false, + "supports_native_streaming": true, + "supports_parallel_function_calling": false, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": false, + "supports_vision": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5-codex": { + "cache_read_input_token_cost": 1.25e-07, + "input_cost_per_token": 1.25e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 1e-05, + "supported_endpoints": [ + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": false, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5.1-codex": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_priority": 2.5e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_priority": 2.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 1e-05, + "output_cost_per_token_priority": 2e-05, + "supported_endpoints": [ + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": false, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5.1-codex-max": { + "cache_read_input_token_cost": 1.25e-07, + "input_cost_per_token": 1.25e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 1e-05, + "supported_endpoints": [ + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": false, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": true + }, + "gpt-5.1-codex-mini": { + "cache_read_input_token_cost": 2.5e-08, + "cache_read_input_token_cost_priority": 4.5e-08, + "input_cost_per_token": 2.5e-07, + "input_cost_per_token_priority": 4.5e-07, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 2e-06, + "output_cost_per_token_priority": 3.6e-06, + "supported_endpoints": [ + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": false, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5.2-codex": { + "cache_read_input_token_cost": 1.75e-07, + "cache_read_input_token_cost_priority": 3.5e-07, + "input_cost_per_token": 1.75e-06, + "input_cost_per_token_priority": 3.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 1.4e-05, + "output_cost_per_token_priority": 2.8e-05, + "supported_endpoints": [ + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": false, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": true + }, + "gpt-5.3-codex": { + "cache_read_input_token_cost": 1.75e-07, + "cache_read_input_token_cost_priority": 3.5e-07, + "input_cost_per_token": 1.75e-06, + "input_cost_per_token_priority": 3.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "responses", + "output_cost_per_token": 1.4e-05, + "output_cost_per_token_priority": 2.8e-05, + "supported_endpoints": [ + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": false, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5-mini": { + "cache_read_input_token_cost": 2.5e-08, + "cache_read_input_token_cost_flex": 1.25e-08, + "cache_read_input_token_cost_priority": 4.5e-08, + "input_cost_per_token": 2.5e-07, + "input_cost_per_token_flex": 1.25e-07, + "input_cost_per_token_priority": 4.5e-07, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 2e-06, + "output_cost_per_token_flex": 1e-06, + "output_cost_per_token_priority": 3.6e-06, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5-mini-2025-08-07": { + "cache_read_input_token_cost": 2.5e-08, + "cache_read_input_token_cost_flex": 1.25e-08, + "cache_read_input_token_cost_priority": 4.5e-08, + "input_cost_per_token": 2.5e-07, + "input_cost_per_token_flex": 1.25e-07, + "input_cost_per_token_priority": 4.5e-07, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 2e-06, + "output_cost_per_token_flex": 1e-06, + "output_cost_per_token_priority": 3.6e-06, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_service_tier": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5-nano": { + "cache_read_input_token_cost": 5e-09, + "cache_read_input_token_cost_flex": 2.5e-09, + "input_cost_per_token": 5e-08, + "input_cost_per_token_flex": 2.5e-08, + "input_cost_per_token_priority": 2.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 4e-07, + "output_cost_per_token_flex": 2e-07, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5-nano-2025-08-07": { + "cache_read_input_token_cost": 5e-09, + "cache_read_input_token_cost_flex": 2.5e-09, + "input_cost_per_token": 5e-08, + "input_cost_per_token_flex": 2.5e-08, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 4e-07, + "output_cost_per_token_flex": 2e-07, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/batch", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-image-1": { + "cache_read_input_image_token_cost": 2.5e-06, + "cache_read_input_token_cost": 1.25e-06, + "input_cost_per_image_token": 1e-05, + "input_cost_per_token": 5e-06, + "litellm_provider": "openai", + "mode": "image_generation", + "output_cost_per_image_token": 4e-05, + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ] + }, + "gpt-image-1-mini": { + "cache_read_input_image_token_cost": 2.5e-07, + "cache_read_input_token_cost": 2e-07, + "input_cost_per_image_token": 2.5e-06, + "input_cost_per_token": 2e-06, + "litellm_provider": "openai", + "mode": "image_generation", + "output_cost_per_image_token": 8e-06, + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ] + }, "gpt-realtime": { "cache_creation_input_audio_token_cost": 4e-07, "cache_read_input_token_cost": 4e-07, @@ -21487,38 +19342,6 @@ "supports_system_messages": true, "supports_tool_choice": true }, - "gpt-realtime-2025-08-28": { - "cache_creation_input_audio_token_cost": 4e-07, - "cache_read_input_token_cost": 4e-07, - "input_cost_per_audio_token": 3.2e-05, - "input_cost_per_image": 5e-06, - "input_cost_per_token": 4e-06, - "litellm_provider": "openai", - "max_input_tokens": 32000, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_audio_token": 6.4e-05, - "output_cost_per_token": 1.6e-05, - "supported_endpoints": [ - "/v1/realtime" - ], - "supported_modalities": [ - "text", - "image", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, "gpt-realtime-mini": { "cache_creation_input_audio_token_cost": 3e-07, "cache_read_input_audio_token_cost": 3e-07, @@ -21550,53 +19373,19 @@ "supports_system_messages": true, "supports_tool_choice": true }, - "gpt-realtime-mini-2025-10-06": { - "cache_creation_input_audio_token_cost": 3e-07, - "cache_read_input_audio_token_cost": 3e-07, - "cache_read_input_token_cost": 6e-08, - "input_cost_per_audio_token": 1e-05, - "input_cost_per_image": 8e-07, - "input_cost_per_token": 6e-07, + "gpt-realtime-2025-08-28": { + "cache_creation_input_audio_token_cost": 4e-07, + "cache_read_input_token_cost": 4e-07, + "input_cost_per_audio_token": 3.2e-05, + "input_cost_per_image": 5e-06, + "input_cost_per_token": 4e-06, "litellm_provider": "openai", - "max_input_tokens": 128000, + "max_input_tokens": 32000, "max_output_tokens": 4096, "max_tokens": 4096, "mode": "chat", - "output_cost_per_audio_token": 2e-05, - "output_cost_per_token": 2.4e-06, - "supported_endpoints": [ - "/v1/realtime" - ], - "supported_modalities": [ - "text", - "image", - "audio" - ], - "supported_output_modalities": [ - "text", - "audio" - ], - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true - }, - "gpt-realtime-mini-2025-12-15": { - "cache_creation_input_audio_token_cost": 3e-07, - "cache_read_input_audio_token_cost": 3e-07, - "cache_read_input_token_cost": 6e-08, - "input_cost_per_audio_token": 1e-05, - "input_cost_per_image": 8e-07, - "input_cost_per_token": 6e-07, - "litellm_provider": "openai", - "max_input_tokens": 128000, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_audio_token": 2e-05, - "output_cost_per_token": 2.4e-06, + "output_cost_per_audio_token": 6.4e-05, + "output_cost_per_token": 1.6e-05, "supported_endpoints": [ "/v1/realtime" ], @@ -21792,18 +19581,120 @@ ], "supports_tool_choice": false }, - "groq/gemma-7b-it": { - "input_cost_per_token": 5e-08, - "litellm_provider": "groq", - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, + "lemonade/Qwen3-Coder-30B-A3B-Instruct-GGUF": { + "input_cost_per_token": 0, + "litellm_provider": "lemonade", + "max_tokens": 32768, + "max_input_tokens": 262144, + "max_output_tokens": 32768, "mode": "chat", - "output_cost_per_token": 8e-08, + "output_cost_per_token": 0, "supports_function_calling": true, - "supports_response_schema": false, + "supports_response_schema": true, "supports_tool_choice": true }, + "lemonade/gpt-oss-20b-mxfp4-GGUF": { + "input_cost_per_token": 0, + "litellm_provider": "lemonade", + "max_tokens": 32768, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "mode": "chat", + "output_cost_per_token": 0, + "supports_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "lemonade/gpt-oss-120b-mxfp-GGUF": { + "input_cost_per_token": 0, + "litellm_provider": "lemonade", + "max_tokens": 32768, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "mode": "chat", + "output_cost_per_token": 0, + "supports_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "lemonade/Gemma-3-4b-it-GGUF": { + "input_cost_per_token": 0, + "litellm_provider": "lemonade", + "max_tokens": 8192, + "max_input_tokens": 128000, + "max_output_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 0, + "supports_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "lemonade/Qwen3-4B-Instruct-2507-GGUF": { + "input_cost_per_token": 0, + "litellm_provider": "lemonade", + "max_tokens": 32768, + "max_input_tokens": 262144, + "max_output_tokens": 32768, + "mode": "chat", + "output_cost_per_token": 0, + "supports_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "amazon-nova/nova-micro-v1": { + "input_cost_per_token": 3.5e-08, + "litellm_provider": "amazon_nova", + "max_input_tokens": 128000, + "max_output_tokens": 10000, + "max_tokens": 10000, + "mode": "chat", + "output_cost_per_token": 1.4e-07, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_response_schema": true + }, + "amazon-nova/nova-lite-v1": { + "input_cost_per_token": 6e-08, + "litellm_provider": "amazon_nova", + "max_input_tokens": 300000, + "max_output_tokens": 10000, + "max_tokens": 10000, + "mode": "chat", + "output_cost_per_token": 2.4e-07, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_vision": true + }, + "amazon-nova/nova-premier-v1": { + "input_cost_per_token": 2.5e-06, + "litellm_provider": "amazon_nova", + "max_input_tokens": 1000000, + "max_output_tokens": 10000, + "max_tokens": 10000, + "mode": "chat", + "output_cost_per_token": 1.25e-05, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": false, + "supports_response_schema": true, + "supports_vision": true + }, + "amazon-nova/nova-pro-v1": { + "input_cost_per_token": 8e-07, + "litellm_provider": "amazon_nova", + "max_input_tokens": 300000, + "max_output_tokens": 10000, + "max_tokens": 10000, + "mode": "chat", + "output_cost_per_token": 3.2e-06, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_vision": true + }, "groq/llama-3.1-8b-instant": { "input_cost_per_token": 5e-08, "litellm_provider": "groq", @@ -21828,6 +19719,27 @@ "supports_response_schema": false, "supports_tool_choice": true }, + "groq/gemma-7b-it": { + "input_cost_per_token": 5e-08, + "litellm_provider": "groq", + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 8e-08, + "supports_function_calling": true, + "supports_response_schema": false, + "supports_tool_choice": true + }, + "groq/meta-llama/llama-guard-4-12b": { + "input_cost_per_token": 2e-07, + "litellm_provider": "groq", + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 2e-07 + }, "groq/meta-llama/llama-4-maverick-17b-128e-instruct": { "input_cost_per_token": 2e-07, "litellm_provider": "groq", @@ -21854,15 +19766,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "groq/meta-llama/llama-guard-4-12b": { - "input_cost_per_token": 2e-07, - "litellm_provider": "groq", - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 2e-07 - }, "groq/moonshotai/kimi-k2-instruct-0905": { "input_cost_per_token": 1e-06, "output_cost_per_token": 3e-06, @@ -22016,28 +19919,6 @@ "/v1/images/generations" ] }, - "high/1024-x-1024/gpt-image-1.5": { - "input_cost_per_image": 0.133, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "high/1024-x-1024/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.133, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, "high/1024-x-1536/gpt-image-1": { "input_cost_per_image": 0.25, "input_cost_per_pixel": 1.58945719e-07, @@ -22048,28 +19929,6 @@ "/v1/images/generations" ] }, - "high/1024-x-1536/gpt-image-1.5": { - "input_cost_per_image": 0.2, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "high/1024-x-1536/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.2, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, "high/1536-x-1024/gpt-image-1": { "input_cost_per_image": 0.25, "input_cost_per_pixel": 1.58945719e-07, @@ -22080,28 +19939,6 @@ "/v1/images/generations" ] }, - "high/1536-x-1024/gpt-image-1.5": { - "input_cost_per_image": 0.2, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "high/1536-x-1024/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.2, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, "hyperbolic/NousResearch/Hermes-3-Llama-3.1-70B": { "input_cost_per_token": 1.2e-07, "litellm_provider": "hyperbolic", @@ -22437,28 +20274,6 @@ "mode": "rerank", "output_cost_per_token": 1.8e-08 }, - "jp.anthropic.claude-haiku-4-5-20251001-v1:0": { - "cache_creation_input_token_cost": 1.375e-06, - "cache_read_input_token_cost": 1.1e-07, - "input_cost_per_token": 1.1e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 5.5e-06, - "source": "https://aws.amazon.com/about-aws/whats-new/2025/10/claude-4-5-haiku-anthropic-amazon-bedrock", - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, "jp.anthropic.claude-sonnet-4-5-20250929-v1:0": { "cache_creation_input_token_cost": 4.125e-06, "cache_read_input_token_cost": 3.3e-07, @@ -22489,19 +20304,27 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 346 }, - "kimi-k2-thinking-251104": { - "input_cost_per_token": 0.0, - "litellm_provider": "volcengine", - "max_input_tokens": 229376, - "max_output_tokens": 32768, - "max_tokens": 32768, + "jp.anthropic.claude-haiku-4-5-20251001-v1:0": { + "cache_creation_input_token_cost": 1.375e-06, + "cache_read_input_token_cost": 1.1e-07, + "input_cost_per_token": 1.1e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, "mode": "chat", - "output_cost_per_token": 0.0, + "output_cost_per_token": 5.5e-06, + "source": "https://aws.amazon.com/about-aws/whats-new/2025/10/claude-4-5-haiku-anthropic-amazon-bedrock", "supports_assistant_prefill": true, + "supports_computer_use": true, "supports_function_calling": true, + "supports_pdf_input": true, "supports_prompt_caching": true, "supports_reasoning": true, - "supports_tool_choice": true + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 }, "lambda_ai/deepseek-llama3.3-70b": { "input_cost_per_token": 2e-07, @@ -22768,251 +20591,6 @@ "supports_system_messages": true, "supports_tool_choice": true }, - "lemonade/Gemma-3-4b-it-GGUF": { - "input_cost_per_token": 0, - "litellm_provider": "lemonade", - "max_tokens": 8192, - "max_input_tokens": 128000, - "max_output_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 0, - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "lemonade/Qwen3-4B-Instruct-2507-GGUF": { - "input_cost_per_token": 0, - "litellm_provider": "lemonade", - "max_tokens": 32768, - "max_input_tokens": 262144, - "max_output_tokens": 32768, - "mode": "chat", - "output_cost_per_token": 0, - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "lemonade/Qwen3-Coder-30B-A3B-Instruct-GGUF": { - "input_cost_per_token": 0, - "litellm_provider": "lemonade", - "max_tokens": 32768, - "max_input_tokens": 262144, - "max_output_tokens": 32768, - "mode": "chat", - "output_cost_per_token": 0, - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "lemonade/gpt-oss-120b-mxfp-GGUF": { - "input_cost_per_token": 0, - "litellm_provider": "lemonade", - "max_tokens": 32768, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "mode": "chat", - "output_cost_per_token": 0, - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "lemonade/gpt-oss-20b-mxfp4-GGUF": { - "input_cost_per_token": 0, - "litellm_provider": "lemonade", - "max_tokens": 32768, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "mode": "chat", - "output_cost_per_token": 0, - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "linkup/search": { - "input_cost_per_query": 0.00587, - "litellm_provider": "linkup", - "mode": "search" - }, - "linkup/search-deep": { - "input_cost_per_query": 0.05867, - "litellm_provider": "linkup", - "mode": "search" - }, - "llamagate/codellama-7b": { - "max_tokens": 4096, - "max_input_tokens": 16384, - "max_output_tokens": 4096, - "input_cost_per_token": 6e-08, - "output_cost_per_token": 1.2e-07, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true - }, - "llamagate/deepseek-coder-6.7b": { - "max_tokens": 4096, - "max_input_tokens": 16384, - "max_output_tokens": 4096, - "input_cost_per_token": 6e-08, - "output_cost_per_token": 1.2e-07, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true - }, - "llamagate/deepseek-r1-7b-qwen": { - "max_tokens": 16384, - "max_input_tokens": 131072, - "max_output_tokens": 16384, - "input_cost_per_token": 8e-08, - "output_cost_per_token": 1.5e-07, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true, - "supports_reasoning": true - }, - "llamagate/deepseek-r1-8b": { - "max_tokens": 16384, - "max_input_tokens": 65536, - "max_output_tokens": 16384, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true, - "supports_reasoning": true - }, - "llamagate/dolphin3-8b": { - "max_tokens": 8192, - "max_input_tokens": 128000, - "max_output_tokens": 8192, - "input_cost_per_token": 8e-08, - "output_cost_per_token": 1.5e-07, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true - }, - "llamagate/gemma3-4b": { - "max_tokens": 8192, - "max_input_tokens": 128000, - "max_output_tokens": 8192, - "input_cost_per_token": 3e-08, - "output_cost_per_token": 8e-08, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true, - "supports_vision": true - }, - "llamagate/llama-3.1-8b": { - "max_tokens": 8192, - "max_input_tokens": 131072, - "max_output_tokens": 8192, - "input_cost_per_token": 3e-08, - "output_cost_per_token": 5e-08, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true - }, - "llamagate/llama-3.2-3b": { - "max_tokens": 8192, - "max_input_tokens": 131072, - "max_output_tokens": 8192, - "input_cost_per_token": 4e-08, - "output_cost_per_token": 8e-08, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true - }, - "llamagate/llava-7b": { - "max_tokens": 2048, - "max_input_tokens": 4096, - "max_output_tokens": 2048, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 2e-07, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_response_schema": true, - "supports_vision": true - }, - "llamagate/mistral-7b-v0.3": { - "max_tokens": 8192, - "max_input_tokens": 32768, - "max_output_tokens": 8192, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1.5e-07, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true - }, - "llamagate/nomic-embed-text": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "input_cost_per_token": 2e-08, - "output_cost_per_token": 0, - "litellm_provider": "llamagate", - "mode": "embedding" - }, - "llamagate/openthinker-7b": { - "max_tokens": 8192, - "max_input_tokens": 32768, - "max_output_tokens": 8192, - "input_cost_per_token": 8e-08, - "output_cost_per_token": 1.5e-07, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true, - "supports_reasoning": true - }, - "llamagate/qwen2.5-coder-7b": { - "max_tokens": 8192, - "max_input_tokens": 32768, - "max_output_tokens": 8192, - "input_cost_per_token": 6e-08, - "output_cost_per_token": 1.2e-07, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true - }, - "llamagate/qwen3-8b": { - "max_tokens": 8192, - "max_input_tokens": 32768, - "max_output_tokens": 8192, - "input_cost_per_token": 4e-08, - "output_cost_per_token": 1.4e-07, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true - }, - "llamagate/qwen3-embedding-8b": { - "max_tokens": 40960, - "max_input_tokens": 40960, - "input_cost_per_token": 2e-08, - "output_cost_per_token": 0, - "litellm_provider": "llamagate", - "mode": "embedding" - }, - "llamagate/qwen3-vl-8b": { - "max_tokens": 8192, - "max_input_tokens": 32768, - "max_output_tokens": 8192, - "input_cost_per_token": 1.5e-07, - "output_cost_per_token": 5.5e-07, - "litellm_provider": "llamagate", - "mode": "chat", - "supports_function_calling": true, - "supports_response_schema": true, - "supports_vision": true - }, "low/1024-x-1024/gpt-image-1": { "input_cost_per_image": 0.011, "input_cost_per_pixel": 1.0490417e-08, @@ -23023,36 +20601,6 @@ "/v1/images/generations" ] }, - "low/1024-x-1024/gpt-image-1-mini": { - "input_cost_per_image": 0.005, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations" - ] - }, - "low/1024-x-1024/gpt-image-1.5": { - "input_cost_per_image": 0.009, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "low/1024-x-1024/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.009, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, "low/1024-x-1536/gpt-image-1": { "input_cost_per_image": 0.016, "input_cost_per_pixel": 1.0172526e-08, @@ -23063,36 +20611,6 @@ "/v1/images/generations" ] }, - "low/1024-x-1536/gpt-image-1-mini": { - "input_cost_per_image": 0.006, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations" - ] - }, - "low/1024-x-1536/gpt-image-1.5": { - "input_cost_per_image": 0.013, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "low/1024-x-1536/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.013, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, "low/1536-x-1024/gpt-image-1": { "input_cost_per_image": 0.016, "input_cost_per_pixel": 1.0172526e-08, @@ -23103,36 +20621,6 @@ "/v1/images/generations" ] }, - "low/1536-x-1024/gpt-image-1-mini": { - "input_cost_per_image": 0.006, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations" - ] - }, - "low/1536-x-1024/gpt-image-1.5": { - "input_cost_per_image": 0.013, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "low/1536-x-1024/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.013, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, "luminous-base": { "input_cost_per_token": 3e-05, "litellm_provider": "aleph_alpha", @@ -23199,36 +20687,6 @@ "/v1/images/generations" ] }, - "medium/1024-x-1024/gpt-image-1-mini": { - "input_cost_per_image": 0.011, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations" - ] - }, - "medium/1024-x-1024/gpt-image-1.5": { - "input_cost_per_image": 0.034, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "medium/1024-x-1024/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.034, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, "medium/1024-x-1536/gpt-image-1": { "input_cost_per_image": 0.063, "input_cost_per_pixel": 4.0054321e-08, @@ -23239,6 +20697,48 @@ "/v1/images/generations" ] }, + "medium/1536-x-1024/gpt-image-1": { + "input_cost_per_image": 0.063, + "input_cost_per_pixel": 4.0054321e-08, + "litellm_provider": "openai", + "mode": "image_generation", + "output_cost_per_pixel": 0.0, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "low/1024-x-1024/gpt-image-1-mini": { + "input_cost_per_image": 0.005, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "low/1024-x-1536/gpt-image-1-mini": { + "input_cost_per_image": 0.006, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "low/1536-x-1024/gpt-image-1-mini": { + "input_cost_per_image": 0.006, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "medium/1024-x-1024/gpt-image-1-mini": { + "input_cost_per_image": 0.011, + "litellm_provider": "openai", + "mode": "image_generation", + "supported_endpoints": [ + "/v1/images/generations" + ] + }, "medium/1024-x-1536/gpt-image-1-mini": { "input_cost_per_image": 0.015, "litellm_provider": "openai", @@ -23247,38 +20747,6 @@ "/v1/images/generations" ] }, - "medium/1024-x-1536/gpt-image-1.5": { - "input_cost_per_image": 0.05, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "medium/1024-x-1536/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.05, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "medium/1536-x-1024/gpt-image-1": { - "input_cost_per_image": 0.063, - "input_cost_per_pixel": 4.0054321e-08, - "litellm_provider": "openai", - "mode": "image_generation", - "output_cost_per_pixel": 0.0, - "supported_endpoints": [ - "/v1/images/generations" - ] - }, "medium/1536-x-1024/gpt-image-1-mini": { "input_cost_per_image": 0.015, "litellm_provider": "openai", @@ -23287,28 +20755,6 @@ "/v1/images/generations" ] }, - "medium/1536-x-1024/gpt-image-1.5": { - "input_cost_per_image": 0.05, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "medium/1536-x-1024/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.05, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, "medlm-large": { "input_cost_per_character": 5e-06, "litellm_provider": "vertex_ai-language-models", @@ -23588,20 +21034,37 @@ "supports_tool_choice": true, "source": "https://aws.amazon.com/bedrock/pricing/" }, - "minimax/MiniMax-M2": { - "input_cost_per_token": 3e-07, - "output_cost_per_token": 1.2e-06, - "cache_read_input_token_cost": 3e-08, - "cache_creation_input_token_cost": 3.75e-07, + "minimax/speech-02-hd": { + "input_cost_per_character": 0.0001, "litellm_provider": "minimax", - "mode": "chat", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_system_messages": true, - "max_input_tokens": 200000, - "max_output_tokens": 8192 + "mode": "audio_speech", + "supported_endpoints": [ + "/v1/audio/speech" + ] + }, + "minimax/speech-02-turbo": { + "input_cost_per_character": 6e-05, + "litellm_provider": "minimax", + "mode": "audio_speech", + "supported_endpoints": [ + "/v1/audio/speech" + ] + }, + "minimax/speech-2.6-hd": { + "input_cost_per_character": 0.0001, + "litellm_provider": "minimax", + "mode": "audio_speech", + "supported_endpoints": [ + "/v1/audio/speech" + ] + }, + "minimax/speech-2.6-turbo": { + "input_cost_per_character": 6e-05, + "litellm_provider": "minimax", + "mode": "audio_speech", + "supported_endpoints": [ + "/v1/audio/speech" + ] }, "minimax/MiniMax-M2.1": { "input_cost_per_token": 3e-07, @@ -23663,37 +21126,20 @@ "max_input_tokens": 1000000, "max_output_tokens": 8192 }, - "minimax/speech-02-hd": { - "input_cost_per_character": 0.0001, + "minimax/MiniMax-M2": { + "input_cost_per_token": 3e-07, + "output_cost_per_token": 1.2e-06, + "cache_read_input_token_cost": 3e-08, + "cache_creation_input_token_cost": 3.75e-07, "litellm_provider": "minimax", - "mode": "audio_speech", - "supported_endpoints": [ - "/v1/audio/speech" - ] - }, - "minimax/speech-02-turbo": { - "input_cost_per_character": 6e-05, - "litellm_provider": "minimax", - "mode": "audio_speech", - "supported_endpoints": [ - "/v1/audio/speech" - ] - }, - "minimax/speech-2.6-hd": { - "input_cost_per_character": 0.0001, - "litellm_provider": "minimax", - "mode": "audio_speech", - "supported_endpoints": [ - "/v1/audio/speech" - ] - }, - "minimax/speech-2.6-turbo": { - "input_cost_per_character": 6e-05, - "litellm_provider": "minimax", - "mode": "audio_speech", - "supported_endpoints": [ - "/v1/audio/speech" - ] + "mode": "chat", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_system_messages": true, + "max_input_tokens": 200000, + "max_output_tokens": 8192 }, "mistral.devstral-2-123b": { "input_cost_per_token": 4e-07, @@ -23863,20 +21309,6 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "mistral/codestral-embed": { - "input_cost_per_token": 1.5e-07, - "litellm_provider": "mistral", - "max_input_tokens": 8192, - "max_tokens": 8192, - "mode": "embedding" - }, - "mistral/codestral-embed-2505": { - "input_cost_per_token": 1.5e-07, - "litellm_provider": "mistral", - "max_input_tokens": 8192, - "max_tokens": 8192, - "mode": "embedding" - }, "mistral/codestral-latest": { "input_cost_per_token": 1e-06, "litellm_provider": "mistral", @@ -23901,34 +21333,6 @@ "supports_assistant_prefill": true, "supports_tool_choice": true }, - "mistral/devstral-2512": { - "input_cost_per_token": 4e-07, - "litellm_provider": "mistral", - "max_input_tokens": 256000, - "max_output_tokens": 256000, - "max_tokens": 256000, - "mode": "chat", - "output_cost_per_token": 2e-06, - "source": "https://mistral.ai/news/devstral-2-vibe-cli", - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "mistral/devstral-latest": { - "input_cost_per_token": 4e-07, - "litellm_provider": "mistral", - "max_input_tokens": 256000, - "max_output_tokens": 256000, - "max_tokens": 256000, - "mode": "chat", - "output_cost_per_token": 2e-06, - "source": "https://mistral.ai/news/devstral-2-vibe-cli", - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, "mistral/devstral-medium-2507": { "input_cost_per_token": 4e-07, "litellm_provider": "mistral", @@ -23943,20 +21347,6 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "mistral/devstral-medium-latest": { - "input_cost_per_token": 4e-07, - "litellm_provider": "mistral", - "max_input_tokens": 256000, - "max_output_tokens": 256000, - "max_tokens": 256000, - "mode": "chat", - "output_cost_per_token": 2e-06, - "source": "https://mistral.ai/news/devstral-2-vibe-cli", - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, "mistral/devstral-small-2505": { "input_cost_per_token": 1e-07, "litellm_provider": "mistral", @@ -24013,18 +21403,45 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "mistral/magistral-medium-1-2-2509": { - "input_cost_per_token": 2e-06, + "mistral/devstral-latest": { + "input_cost_per_token": 4e-07, "litellm_provider": "mistral", - "max_input_tokens": 40000, - "max_output_tokens": 40000, - "max_tokens": 40000, + "max_input_tokens": 256000, + "max_output_tokens": 256000, + "max_tokens": 256000, "mode": "chat", - "output_cost_per_token": 5e-06, - "source": "https://mistral.ai/news/magistral", + "output_cost_per_token": 2e-06, + "source": "https://mistral.ai/news/devstral-2-vibe-cli", + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "mistral/devstral-medium-latest": { + "input_cost_per_token": 4e-07, + "litellm_provider": "mistral", + "max_input_tokens": 256000, + "max_output_tokens": 256000, + "max_tokens": 256000, + "mode": "chat", + "output_cost_per_token": 2e-06, + "source": "https://mistral.ai/news/devstral-2-vibe-cli", + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "mistral/devstral-2512": { + "input_cost_per_token": 4e-07, + "litellm_provider": "mistral", + "max_input_tokens": 256000, + "max_output_tokens": 256000, + "max_tokens": 256000, + "mode": "chat", + "output_cost_per_token": 2e-06, + "source": "https://mistral.ai/news/devstral-2-vibe-cli", "supports_assistant_prefill": true, "supports_function_calling": true, - "supports_reasoning": true, "supports_response_schema": true, "supports_tool_choice": true }, @@ -24058,7 +21475,7 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "mistral/magistral-medium-latest": { + "mistral/magistral-medium-1-2-2509": { "input_cost_per_token": 2e-06, "litellm_provider": "mistral", "max_input_tokens": 40000, @@ -24073,15 +21490,35 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "mistral/magistral-small-1-2-2509": { - "input_cost_per_token": 5e-07, + "mistral/mistral-ocr-latest": { + "litellm_provider": "mistral", + "ocr_cost_per_page": 0.001, + "annotation_cost_per_page": 0.003, + "mode": "ocr", + "supported_endpoints": [ + "/v1/ocr" + ], + "source": "https://mistral.ai/pricing#api-pricing" + }, + "mistral/mistral-ocr-2505-completion": { + "litellm_provider": "mistral", + "ocr_cost_per_page": 0.001, + "annotation_cost_per_page": 0.003, + "mode": "ocr", + "supported_endpoints": [ + "/v1/ocr" + ], + "source": "https://mistral.ai/pricing#api-pricing" + }, + "mistral/magistral-medium-latest": { + "input_cost_per_token": 2e-06, "litellm_provider": "mistral", "max_input_tokens": 40000, "max_output_tokens": 40000, "max_tokens": 40000, "mode": "chat", - "output_cost_per_token": 1.5e-06, - "source": "https://mistral.ai/pricing#api-pricing", + "output_cost_per_token": 5e-06, + "source": "https://mistral.ai/news/magistral", "supports_assistant_prefill": true, "supports_function_calling": true, "supports_reasoning": true, @@ -24118,50 +21555,20 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "mistral/ministral-3-14b-2512": { - "input_cost_per_token": 2e-07, + "mistral/magistral-small-1-2-2509": { + "input_cost_per_token": 5e-07, "litellm_provider": "mistral", - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, + "max_input_tokens": 40000, + "max_output_tokens": 40000, + "max_tokens": 40000, "mode": "chat", - "output_cost_per_token": 2e-07, - "source": "https://mistral.ai/pricing", + "output_cost_per_token": 1.5e-06, + "source": "https://mistral.ai/pricing#api-pricing", "supports_assistant_prefill": true, "supports_function_calling": true, + "supports_reasoning": true, "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true - }, - "mistral/ministral-3-3b-2512": { - "input_cost_per_token": 1e-07, - "litellm_provider": "mistral", - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "max_tokens": 131072, - "mode": "chat", - "output_cost_per_token": 1e-07, - "source": "https://mistral.ai/pricing", - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true - }, - "mistral/ministral-3-8b-2512": { - "input_cost_per_token": 1.5e-07, - "litellm_provider": "mistral", - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 1.5e-07, - "source": "https://mistral.ai/pricing", - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true + "supports_tool_choice": true }, "mistral/mistral-embed": { "input_cost_per_token": 1e-07, @@ -24170,6 +21577,20 @@ "max_tokens": 8192, "mode": "embedding" }, + "mistral/codestral-embed": { + "input_cost_per_token": 1.5e-07, + "litellm_provider": "mistral", + "max_input_tokens": 8192, + "max_tokens": 8192, + "mode": "embedding" + }, + "mistral/codestral-embed-2505": { + "input_cost_per_token": 1.5e-07, + "litellm_provider": "mistral", + "max_input_tokens": 8192, + "max_tokens": 8192, + "mode": "embedding" + }, "mistral/mistral-large-2402": { "input_cost_per_token": 4e-06, "litellm_provider": "mistral", @@ -24209,7 +21630,7 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "mistral/mistral-large-2512": { + "mistral/mistral-large-latest": { "input_cost_per_token": 5e-07, "litellm_provider": "mistral", "max_input_tokens": 262144, @@ -24239,7 +21660,7 @@ "supports_tool_choice": true, "supports_vision": true }, - "mistral/mistral-large-latest": { + "mistral/mistral-large-2512": { "input_cost_per_token": 5e-07, "litellm_provider": "mistral", "max_input_tokens": 262144, @@ -24291,6 +21712,20 @@ "supports_response_schema": true, "supports_tool_choice": true }, + "mistral/mistral-medium-latest": { + "input_cost_per_token": 4e-07, + "litellm_provider": "mistral", + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "max_tokens": 131072, + "mode": "chat", + "output_cost_per_token": 2e-06, + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true + }, "mistral/mistral-medium-3-1-2508": { "input_cost_per_token": 4e-07, "litellm_provider": "mistral", @@ -24306,40 +21741,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "mistral/mistral-medium-latest": { - "input_cost_per_token": 4e-07, - "litellm_provider": "mistral", - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "max_tokens": 131072, - "mode": "chat", - "output_cost_per_token": 2e-06, - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true - }, - "mistral/mistral-ocr-2505-completion": { - "litellm_provider": "mistral", - "ocr_cost_per_page": 0.001, - "annotation_cost_per_page": 0.003, - "mode": "ocr", - "supported_endpoints": [ - "/v1/ocr" - ], - "source": "https://mistral.ai/pricing#api-pricing" - }, - "mistral/mistral-ocr-latest": { - "litellm_provider": "mistral", - "ocr_cost_per_page": 0.001, - "annotation_cost_per_page": 0.003, - "mode": "ocr", - "supported_endpoints": [ - "/v1/ocr" - ], - "source": "https://mistral.ai/pricing#api-pricing" - }, "mistral/mistral-small": { "input_cost_per_token": 1e-07, "litellm_provider": "mistral", @@ -24353,6 +21754,21 @@ "supports_response_schema": true, "supports_tool_choice": true }, + "mistral/mistral-small-latest": { + "input_cost_per_token": 6e-08, + "litellm_provider": "mistral", + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "max_tokens": 131072, + "mode": "chat", + "output_cost_per_token": 1.8e-07, + "source": "https://mistral.ai/pricing", + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true + }, "mistral/mistral-small-3-2-2506": { "input_cost_per_token": 6e-08, "litellm_provider": "mistral", @@ -24368,14 +21784,44 @@ "supports_tool_choice": true, "supports_vision": true }, - "mistral/mistral-small-latest": { - "input_cost_per_token": 6e-08, + "mistral/ministral-3-3b-2512": { + "input_cost_per_token": 1e-07, "litellm_provider": "mistral", "max_input_tokens": 131072, "max_output_tokens": 131072, "max_tokens": 131072, "mode": "chat", - "output_cost_per_token": 1.8e-07, + "output_cost_per_token": 1e-07, + "source": "https://mistral.ai/pricing", + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "mistral/ministral-3-8b-2512": { + "input_cost_per_token": 1.5e-07, + "litellm_provider": "mistral", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 1.5e-07, + "source": "https://mistral.ai/pricing", + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "mistral/ministral-3-14b-2512": { + "input_cost_per_token": 2e-07, + "litellm_provider": "mistral", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 2e-07, "source": "https://mistral.ai/pricing", "supports_assistant_prefill": true, "supports_function_calling": true, @@ -24524,6 +21970,20 @@ "supports_reasoning": true, "supports_system_messages": true }, + "moonshotai.kimi-k2.5": { + "input_cost_per_token": 6e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 3e-06, + "supports_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, "moonshot/kimi-k2-0711-preview": { "cache_read_input_token_cost": 1.5e-07, "input_cost_per_token": 6e-07, @@ -24552,34 +22012,6 @@ "supports_tool_choice": true, "supports_web_search": true }, - "moonshot/kimi-k2-thinking": { - "cache_read_input_token_cost": 1.5e-07, - "input_cost_per_token": 6e-07, - "litellm_provider": "moonshot", - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 2.5e-06, - "source": "https://platform.moonshot.ai/docs/pricing/chat#generation-model-kimi-k2", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_web_search": true - }, - "moonshot/kimi-k2-thinking-turbo": { - "cache_read_input_token_cost": 1.5e-07, - "input_cost_per_token": 1.15e-06, - "litellm_provider": "moonshot", - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 8e-06, - "source": "https://platform.moonshot.ai/docs/pricing/chat#generation-model-kimi-k2", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_web_search": true - }, "moonshot/kimi-k2-turbo-preview": { "cache_read_input_token_cost": 1.5e-07, "input_cost_per_token": 1.15e-06, @@ -24677,6 +22109,34 @@ "source": "https://platform.moonshot.ai/docs/pricing/chat#generation-model-kimi-k2", "supports_vision": true }, + "moonshot/kimi-k2-thinking": { + "cache_read_input_token_cost": 1.5e-07, + "input_cost_per_token": 6e-07, + "litellm_provider": "moonshot", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 2.5e-06, + "source": "https://platform.moonshot.ai/docs/pricing/chat#generation-model-kimi-k2", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_web_search": true + }, + "moonshot/kimi-k2-thinking-turbo": { + "cache_read_input_token_cost": 1.5e-07, + "input_cost_per_token": 1.15e-06, + "litellm_provider": "moonshot", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 8e-06, + "source": "https://platform.moonshot.ai/docs/pricing/chat#generation-model-kimi-k2", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_web_search": true + }, "moonshot/moonshot-v1-128k": { "input_cost_per_token": 2e-06, "litellm_provider": "moonshot", @@ -24800,20 +22260,6 @@ "supports_function_calling": true, "supports_tool_choice": true }, - "moonshotai.kimi-k2.5": { - "input_cost_per_token": 6e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 3e-06, - "supports_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, "morph/morph-v3-fast": { "input_cost_per_token": 8e-07, "litellm_provider": "morph", @@ -24888,1463 +22334,6 @@ "video" ] }, - "nebius/BAAI/bge-en-icl": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "input_cost_per_token": 1e-08, - "output_cost_per_token": 0.0, - "litellm_provider": "nebius", - "mode": "embedding", - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/BAAI/bge-multilingual-gemma2": { - "max_tokens": 8192, - "max_input_tokens": 8192, - "input_cost_per_token": 1e-08, - "output_cost_per_token": 0.0, - "litellm_provider": "nebius", - "mode": "embedding", - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/NousResearch/Hermes-3-Llama-3.1-405B": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 1e-06, - "output_cost_per_token": 3e-06, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/Qwen/QwQ-32B": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1.5e-07, - "output_cost_per_token": 4.5e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "supports_reasoning": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/Qwen/Qwen2-VL-72B-Instruct": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1.3e-07, - "output_cost_per_token": 4e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "supports_vision": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/Qwen/Qwen2-VL-7B-Instruct": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 2e-08, - "output_cost_per_token": 6e-08, - "litellm_provider": "nebius", - "mode": "chat", - "supports_vision": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/Qwen/Qwen2.5-32B-Instruct": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 6e-08, - "output_cost_per_token": 2e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/Qwen/Qwen2.5-72B-Instruct": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 1.3e-07, - "output_cost_per_token": 4e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/Qwen/Qwen2.5-Coder-7B": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1e-08, - "output_cost_per_token": 3e-08, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/Qwen/Qwen2.5-VL-72B-Instruct": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1.3e-07, - "output_cost_per_token": 4e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "supports_vision": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/Qwen/Qwen3-14B": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 8e-08, - "output_cost_per_token": 2.4e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/Qwen/Qwen3-235B-A22B": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 2e-07, - "output_cost_per_token": 6e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/Qwen/Qwen3-30B-A3B": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 3e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/Qwen/Qwen3-32B": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 3e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/Qwen/Qwen3-4B": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 8e-08, - "output_cost_per_token": 2.4e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/deepseek-ai/DeepSeek-R1": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 8e-07, - "output_cost_per_token": 2.4e-06, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "supports_reasoning": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/deepseek-ai/DeepSeek-R1-0528": { - "max_tokens": 164000, - "max_input_tokens": 164000, - "max_output_tokens": 164000, - "input_cost_per_token": 8e-07, - "output_cost_per_token": 2.4e-06, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "supports_reasoning": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/deepseek-ai/DeepSeek-R1-Distill-Llama-70B": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 2.5e-07, - "output_cost_per_token": 7.5e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/deepseek-ai/DeepSeek-V3": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 1.5e-06, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/deepseek-ai/DeepSeek-V3-0324": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 5e-07, - "output_cost_per_token": 1.5e-06, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/google/gemma-3-27b-it": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 6e-08, - "output_cost_per_token": 2e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "supports_vision": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/intfloat/e5-mistral-7b-instruct": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "input_cost_per_token": 1e-08, - "output_cost_per_token": 0.0, - "litellm_provider": "nebius", - "mode": "embedding", - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/meta-llama/Llama-3.3-70B-Instruct": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 1.3e-07, - "output_cost_per_token": 4e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/meta-llama/Llama-Guard-3-8B": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 2e-08, - "output_cost_per_token": 6e-08, - "litellm_provider": "nebius", - "mode": "chat", - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/meta-llama/Meta-Llama-3.1-405B-Instruct": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 1e-06, - "output_cost_per_token": 3e-06, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/meta-llama/Meta-Llama-3.1-70B-Instruct": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 1.3e-07, - "output_cost_per_token": 4e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/meta-llama/Meta-Llama-3.1-8B-Instruct": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 2e-08, - "output_cost_per_token": 6e-08, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/mistralai/Mistral-Nemo-Instruct-2407": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 4e-08, - "output_cost_per_token": 1.2e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/nvidia/Llama-3.1-Nemotron-Ultra-253B-v1": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 6e-07, - "output_cost_per_token": 1.8e-06, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "nebius/nvidia/Llama-3.3-Nemotron-Super-49B-v1": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 1e-07, - "output_cost_per_token": 4e-07, - "litellm_provider": "nebius", - "mode": "chat", - "supports_function_calling": true, - "source": "https://nebius.com/prices-ai-studio" - }, - "novita/Sao10K/L3-8B-Stheno-v3.2": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 5e-08, - "output_cost_per_token": 5e-08, - "max_input_tokens": 8192, - "max_output_tokens": 32000, - "max_tokens": 32000, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true - }, - "novita/baai/bge-m3": { - "litellm_provider": "novita", - "mode": "embedding", - "input_cost_per_token": 1e-08, - "output_cost_per_token": 1e-08, - "max_input_tokens": 8192, - "max_output_tokens": 96000, - "max_tokens": 96000 - }, - "novita/baai/bge-reranker-v2-m3": { - "litellm_provider": "novita", - "mode": "rerank", - "input_cost_per_token": 1e-08, - "output_cost_per_token": 1e-08, - "max_input_tokens": 8000, - "max_output_tokens": 8000, - "max_tokens": 8000 - }, - "novita/baichuan/baichuan-m2-32b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 7e-08, - "output_cost_per_token": 7e-08, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "max_tokens": 131072, - "supports_system_messages": true - }, - "novita/baidu/ernie-4.5-21B-a3b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 7e-08, - "output_cost_per_token": 2.8e-07, - "max_input_tokens": 120000, - "max_output_tokens": 8000, - "max_tokens": 8000, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true - }, - "novita/baidu/ernie-4.5-21B-a3b-thinking": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 7e-08, - "output_cost_per_token": 2.8e-07, - "max_input_tokens": 131072, - "max_output_tokens": 65536, - "max_tokens": 65536, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/baidu/ernie-4.5-300b-a47b-paddle": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2.8e-07, - "output_cost_per_token": 1.1e-06, - "max_input_tokens": 123000, - "max_output_tokens": 12000, - "max_tokens": 12000, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/baidu/ernie-4.5-vl-28b-a3b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1.4e-07, - "output_cost_per_token": 5.6e-07, - "max_input_tokens": 30000, - "max_output_tokens": 8000, - "max_tokens": 8000, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/baidu/ernie-4.5-vl-28b-a3b-thinking": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3.9e-07, - "output_cost_per_token": 3.9e-07, - "max_input_tokens": 131072, - "max_output_tokens": 65536, - "max_tokens": 65536, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true, - "supports_reasoning": true - }, - "novita/baidu/ernie-4.5-vl-424b-a47b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 4.2e-07, - "output_cost_per_token": 1.25e-06, - "max_input_tokens": 123000, - "max_output_tokens": 16000, - "max_tokens": 16000, - "supports_vision": true, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/deepseek/deepseek-ocr": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3e-08, - "output_cost_per_token": 3e-08, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/deepseek/deepseek-prover-v2-671b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 7e-07, - "output_cost_per_token": 2.5e-06, - "max_input_tokens": 160000, - "max_output_tokens": 160000, - "max_tokens": 160000, - "supports_system_messages": true - }, - "novita/deepseek/deepseek-r1-0528": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 7e-07, - "output_cost_per_token": 2.5e-06, - "max_input_tokens": 163840, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "cache_read_input_token_cost": 3.5e-07, - "input_cost_per_token_cache_hit": 3.5e-07, - "supports_reasoning": true - }, - "novita/deepseek/deepseek-r1-0528-qwen3-8b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 6e-08, - "output_cost_per_token": 9e-08, - "max_input_tokens": 128000, - "max_output_tokens": 32000, - "max_tokens": 32000, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/deepseek/deepseek-r1-distill-llama-70b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 8e-07, - "output_cost_per_token": 8e-07, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "supports_reasoning": true - }, - "novita/deepseek/deepseek-r1-distill-qwen-14b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1.5e-07, - "output_cost_per_token": 1.5e-07, - "max_input_tokens": 32768, - "max_output_tokens": 16384, - "max_tokens": 16384, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "supports_reasoning": true - }, - "novita/deepseek/deepseek-r1-distill-qwen-32b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3e-07, - "output_cost_per_token": 3e-07, - "max_input_tokens": 64000, - "max_output_tokens": 32000, - "max_tokens": 32000, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "supports_reasoning": true - }, - "novita/deepseek/deepseek-r1-turbo": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 7e-07, - "output_cost_per_token": 2.5e-06, - "max_input_tokens": 64000, - "max_output_tokens": 16000, - "max_tokens": 16000, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/deepseek/deepseek-v3-0324": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2.7e-07, - "output_cost_per_token": 1.12e-06, - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "max_tokens": 163840, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "cache_read_input_token_cost": 1.35e-07, - "input_cost_per_token_cache_hit": 1.35e-07 - }, - "novita/deepseek/deepseek-v3-turbo": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 4e-07, - "output_cost_per_token": 1.3e-06, - "max_input_tokens": 64000, - "max_output_tokens": 16000, - "max_tokens": 16000, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true - }, - "novita/deepseek/deepseek-v3.1": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2.7e-07, - "output_cost_per_token": 1e-06, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "cache_read_input_token_cost": 1.35e-07, - "input_cost_per_token_cache_hit": 1.35e-07, - "supports_reasoning": true - }, - "novita/deepseek/deepseek-v3.1-terminus": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2.7e-07, - "output_cost_per_token": 1e-06, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "cache_read_input_token_cost": 1.35e-07, - "input_cost_per_token_cache_hit": 1.35e-07, - "supports_reasoning": true - }, - "novita/deepseek/deepseek-v3.2": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2.69e-07, - "output_cost_per_token": 4e-07, - "max_input_tokens": 163840, - "max_output_tokens": 65536, - "max_tokens": 65536, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "cache_read_input_token_cost": 1.345e-07, - "input_cost_per_token_cache_hit": 1.345e-07, - "supports_reasoning": true - }, - "novita/deepseek/deepseek-v3.2-exp": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2.7e-07, - "output_cost_per_token": 4.1e-07, - "max_input_tokens": 163840, - "max_output_tokens": 65536, - "max_tokens": 65536, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "supports_reasoning": true - }, - "novita/google/gemma-3-12b-it": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 5e-08, - "output_cost_per_token": 1e-07, - "max_input_tokens": 131072, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/google/gemma-3-27b-it": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1.19e-07, - "output_cost_per_token": 2e-07, - "max_input_tokens": 98304, - "max_output_tokens": 16384, - "max_tokens": 16384, - "supports_vision": true, - "supports_system_messages": true - }, - "novita/gryphe/mythomax-l2-13b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 9e-08, - "output_cost_per_token": 9e-08, - "max_input_tokens": 4096, - "max_output_tokens": 3200, - "max_tokens": 3200, - "supports_system_messages": true - }, - "novita/kwaipilot/kat-coder-pro": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3e-07, - "output_cost_per_token": 1.2e-06, - "max_input_tokens": 256000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "cache_read_input_token_cost": 6e-08, - "input_cost_per_token_cache_hit": 6e-08 - }, - "novita/meta-llama/llama-3-70b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 5.1e-07, - "output_cost_per_token": 7.4e-07, - "max_input_tokens": 8192, - "max_output_tokens": 8000, - "max_tokens": 8000, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/meta-llama/llama-3-8b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 4e-08, - "output_cost_per_token": 4e-08, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_system_messages": true - }, - "novita/meta-llama/llama-3.1-8b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2e-08, - "output_cost_per_token": 5e-08, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "max_tokens": 16384, - "supports_system_messages": true - }, - "novita/meta-llama/llama-3.2-3b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3e-08, - "output_cost_per_token": 5e-08, - "max_input_tokens": 32768, - "max_output_tokens": 32000, - "max_tokens": 32000, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true - }, - "novita/meta-llama/llama-3.3-70b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1.35e-07, - "output_cost_per_token": 4e-07, - "max_input_tokens": 131072, - "max_output_tokens": 120000, - "max_tokens": 120000, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true - }, - "novita/meta-llama/llama-4-maverick-17b-128e-instruct-fp8": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2.7e-07, - "output_cost_per_token": 8.5e-07, - "max_input_tokens": 1048576, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_vision": true, - "supports_system_messages": true - }, - "novita/meta-llama/llama-4-scout-17b-16e-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1.8e-07, - "output_cost_per_token": 5.9e-07, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "max_tokens": 131072, - "supports_vision": true, - "supports_system_messages": true - }, - "novita/microsoft/wizardlm-2-8x22b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 6.2e-07, - "output_cost_per_token": 6.2e-07, - "max_input_tokens": 65535, - "max_output_tokens": 8000, - "max_tokens": 8000, - "supports_system_messages": true - }, - "novita/minimax/minimax-m2": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3e-07, - "output_cost_per_token": 1.2e-06, - "max_input_tokens": 204800, - "max_output_tokens": 131072, - "max_tokens": 131072, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "cache_read_input_token_cost": 3e-08, - "input_cost_per_token_cache_hit": 3e-08, - "supports_reasoning": true - }, - "novita/minimax/minimax-m2.1": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3e-07, - "output_cost_per_token": 1.2e-06, - "max_input_tokens": 204800, - "max_output_tokens": 131072, - "max_tokens": 131072, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "cache_read_input_token_cost": 3e-08, - "input_cost_per_token_cache_hit": 3e-08 - }, - "novita/minimaxai/minimax-m1-80k": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 5.5e-07, - "output_cost_per_token": 2.2e-06, - "max_input_tokens": 1000000, - "max_output_tokens": 40000, - "max_tokens": 40000, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/mistralai/mistral-nemo": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 4e-08, - "output_cost_per_token": 1.7e-07, - "max_input_tokens": 60288, - "max_output_tokens": 16000, - "max_tokens": 16000, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/moonshotai/kimi-k2-0905": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 6e-07, - "output_cost_per_token": 2.5e-06, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/moonshotai/kimi-k2-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 5.7e-07, - "output_cost_per_token": 2.3e-06, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "max_tokens": 131072, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/moonshotai/kimi-k2-thinking": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 6e-07, - "output_cost_per_token": 2.5e-06, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "supports_reasoning": true - }, - "novita/nousresearch/hermes-2-pro-llama-3-8b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1.4e-07, - "output_cost_per_token": 1.4e-07, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/openai/gpt-oss-120b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 5e-08, - "output_cost_per_token": 2.5e-07, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true, - "supports_reasoning": true - }, - "novita/openai/gpt-oss-20b": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 4e-08, - "output_cost_per_token": 1.5e-07, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true, - "supports_reasoning": true - }, - "novita/paddlepaddle/paddleocr-vl": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2e-08, - "output_cost_per_token": 2e-08, - "max_input_tokens": 16384, - "max_output_tokens": 16384, - "max_tokens": 16384, - "supports_vision": true, - "supports_system_messages": true - }, - "novita/qwen/qwen-2.5-72b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3.8e-07, - "output_cost_per_token": 4e-07, - "max_input_tokens": 32000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/qwen/qwen-mt-plus": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2.5e-07, - "output_cost_per_token": 7.5e-07, - "max_input_tokens": 16384, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_system_messages": true - }, - "novita/qwen/qwen2.5-7b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 7e-08, - "output_cost_per_token": 7e-08, - "max_input_tokens": 32000, - "max_output_tokens": 32000, - "max_tokens": 32000, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/qwen/qwen2.5-vl-72b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 8e-07, - "output_cost_per_token": 8e-07, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_vision": true, - "supports_system_messages": true - }, - "novita/qwen/qwen3-235b-a22b-fp8": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2e-07, - "output_cost_per_token": 8e-07, - "max_input_tokens": 40960, - "max_output_tokens": 20000, - "max_tokens": 20000, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/qwen/qwen3-235b-a22b-instruct-2507": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 9e-08, - "output_cost_per_token": 5.8e-07, - "max_input_tokens": 131072, - "max_output_tokens": 16384, - "max_tokens": 16384, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/qwen/qwen3-235b-a22b-thinking-2507": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3e-07, - "output_cost_per_token": 3e-06, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/qwen/qwen3-30b-a3b-fp8": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 9e-08, - "output_cost_per_token": 4.5e-07, - "max_input_tokens": 40960, - "max_output_tokens": 20000, - "max_tokens": 20000, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/qwen/qwen3-32b-fp8": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1e-07, - "output_cost_per_token": 4.5e-07, - "max_input_tokens": 40960, - "max_output_tokens": 20000, - "max_tokens": 20000, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/qwen/qwen3-4b-fp8": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3e-08, - "output_cost_per_token": 3e-08, - "max_input_tokens": 128000, - "max_output_tokens": 20000, - "max_tokens": 20000, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/qwen/qwen3-8b-fp8": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3.5e-08, - "output_cost_per_token": 1.38e-07, - "max_input_tokens": 128000, - "max_output_tokens": 20000, - "max_tokens": 20000, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/qwen/qwen3-coder-30b-a3b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 7e-08, - "output_cost_per_token": 2.7e-07, - "max_input_tokens": 160000, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/qwen/qwen3-coder-480b-a35b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3e-07, - "output_cost_per_token": 1.3e-06, - "max_input_tokens": 262144, - "max_output_tokens": 65536, - "max_tokens": 65536, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/qwen/qwen3-embedding-0.6b": { - "litellm_provider": "novita", - "mode": "embedding", - "input_cost_per_token": 7e-08, - "output_cost_per_token": 0, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "max_tokens": 32768 - }, - "novita/qwen/qwen3-embedding-8b": { - "litellm_provider": "novita", - "mode": "embedding", - "input_cost_per_token": 7e-08, - "output_cost_per_token": 0, - "max_input_tokens": 32768, - "max_output_tokens": 4096, - "max_tokens": 4096 - }, - "novita/qwen/qwen3-max": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2.11e-06, - "output_cost_per_token": 8.45e-06, - "max_input_tokens": 262144, - "max_output_tokens": 65536, - "max_tokens": 65536, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/qwen/qwen3-next-80b-a3b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1.5e-07, - "output_cost_per_token": 1.5e-06, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/qwen/qwen3-next-80b-a3b-thinking": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1.5e-07, - "output_cost_per_token": 1.5e-06, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "supports_reasoning": true - }, - "novita/qwen/qwen3-omni-30b-a3b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2.5e-07, - "output_cost_per_token": 9.7e-07, - "max_input_tokens": 65536, - "max_output_tokens": 16384, - "max_tokens": 16384, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true, - "supports_audio_input": true, - "supports_audio_output": true - }, - "novita/qwen/qwen3-omni-30b-a3b-thinking": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2.5e-07, - "output_cost_per_token": 9.7e-07, - "max_input_tokens": 65536, - "max_output_tokens": 16384, - "max_tokens": 16384, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true, - "supports_reasoning": true, - "supports_audio_input": true - }, - "novita/qwen/qwen3-reranker-8b": { - "litellm_provider": "novita", - "mode": "rerank", - "input_cost_per_token": 5e-08, - "output_cost_per_token": 5e-08, - "max_input_tokens": 32768, - "max_output_tokens": 4096, - "max_tokens": 4096 - }, - "novita/qwen/qwen3-vl-235b-a22b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3e-07, - "output_cost_per_token": 1.5e-06, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/qwen/qwen3-vl-235b-a22b-thinking": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 9.8e-07, - "output_cost_per_token": 3.95e-06, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_vision": true, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/qwen/qwen3-vl-30b-a3b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2e-07, - "output_cost_per_token": 7e-07, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/qwen/qwen3-vl-30b-a3b-thinking": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2e-07, - "output_cost_per_token": 1e-06, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/qwen/qwen3-vl-8b-instruct": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 8e-08, - "output_cost_per_token": 5e-07, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/sao10k/l3-70b-euryale-v2.1": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1.48e-06, - "output_cost_per_token": 1.48e-06, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true - }, - "novita/sao10k/l3-8b-lunaris": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 5e-08, - "output_cost_per_token": 5e-08, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/sao10k/l31-70b-euryale-v2.2": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1.48e-06, - "output_cost_per_token": 1.48e-06, - "max_input_tokens": 8192, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true - }, - "novita/skywork/r1v4-lite": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 2e-07, - "output_cost_per_token": 6e-07, - "max_input_tokens": 262144, - "max_output_tokens": 65536, - "max_tokens": 65536, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true - }, - "novita/xiaomimimo/mimo-v2-flash": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1e-07, - "output_cost_per_token": 3e-07, - "max_input_tokens": 262144, - "max_output_tokens": 32000, - "max_tokens": 32000, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "cache_read_input_token_cost": 2e-08, - "input_cost_per_token_cache_hit": 2e-08, - "supports_reasoning": true - }, - "novita/zai-org/autoglm-phone-9b-multilingual": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3.5e-08, - "output_cost_per_token": 1.38e-07, - "max_input_tokens": 65536, - "max_output_tokens": 65536, - "max_tokens": 65536, - "supports_vision": true, - "supports_system_messages": true - }, - "novita/zai-org/glm-4.5": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 6e-07, - "output_cost_per_token": 2.2e-06, - "max_input_tokens": 131072, - "max_output_tokens": 98304, - "max_tokens": 98304, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "cache_read_input_token_cost": 1.1e-07, - "input_cost_per_token_cache_hit": 1.1e-07, - "supports_reasoning": true - }, - "novita/zai-org/glm-4.5-air": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 1.3e-07, - "output_cost_per_token": 8.5e-07, - "max_input_tokens": 131072, - "max_output_tokens": 98304, - "max_tokens": 98304, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_reasoning": true - }, - "novita/zai-org/glm-4.5v": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 6e-07, - "output_cost_per_token": 1.8e-06, - "max_input_tokens": 65536, - "max_output_tokens": 16384, - "max_tokens": 16384, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true, - "cache_read_input_token_cost": 1.1e-07, - "input_cost_per_token_cache_hit": 1.1e-07, - "supports_reasoning": true - }, - "novita/zai-org/glm-4.6": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 5.5e-07, - "output_cost_per_token": 2.2e-06, - "max_input_tokens": 204800, - "max_output_tokens": 131072, - "max_tokens": 131072, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "cache_read_input_token_cost": 1.1e-07, - "input_cost_per_token_cache_hit": 1.1e-07, - "supports_reasoning": true - }, - "novita/zai-org/glm-4.6v": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 3e-07, - "output_cost_per_token": 9e-07, - "max_input_tokens": 131072, - "max_output_tokens": 32768, - "max_tokens": 32768, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_response_schema": true, - "cache_read_input_token_cost": 5.5e-08, - "input_cost_per_token_cache_hit": 5.5e-08, - "supports_reasoning": true - }, - "novita/zai-org/glm-4.7": { - "litellm_provider": "novita", - "mode": "chat", - "input_cost_per_token": 6e-07, - "output_cost_per_token": 2.2e-06, - "max_input_tokens": 204800, - "max_output_tokens": 131072, - "max_tokens": 131072, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true, - "supports_system_messages": true, - "supports_response_schema": true, - "cache_read_input_token_cost": 1.1e-07, - "input_cost_per_token_cache_hit": 1.1e-07, - "supports_reasoning": true - }, "nscale/Qwen/QwQ-32B": { "input_cost_per_token": 1.8e-07, "litellm_provider": "nscale", @@ -26490,6 +22479,335 @@ "/v1/images/generations" ] }, + "nebius/deepseek-ai/DeepSeek-R1": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 8e-07, + "output_cost_per_token": 2.4e-06, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "supports_reasoning": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/deepseek-ai/DeepSeek-R1-0528": { + "max_tokens": 164000, + "max_input_tokens": 164000, + "max_output_tokens": 164000, + "input_cost_per_token": 8e-07, + "output_cost_per_token": 2.4e-06, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "supports_reasoning": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/deepseek-ai/DeepSeek-R1-Distill-Llama-70B": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 2.5e-07, + "output_cost_per_token": 7.5e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/deepseek-ai/DeepSeek-V3": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 1.5e-06, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/deepseek-ai/DeepSeek-V3-0324": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 1.5e-06, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/google/gemma-3-27b-it": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 6e-08, + "output_cost_per_token": 2e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "supports_vision": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/meta-llama/Llama-3.3-70B-Instruct": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 1.3e-07, + "output_cost_per_token": 4e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/meta-llama/Llama-Guard-3-8B": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 2e-08, + "output_cost_per_token": 6e-08, + "litellm_provider": "nebius", + "mode": "chat", + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/meta-llama/Meta-Llama-3.1-8B-Instruct": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 2e-08, + "output_cost_per_token": 6e-08, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/meta-llama/Meta-Llama-3.1-70B-Instruct": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 1.3e-07, + "output_cost_per_token": 4e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/meta-llama/Meta-Llama-3.1-405B-Instruct": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 1e-06, + "output_cost_per_token": 3e-06, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/mistralai/Mistral-Nemo-Instruct-2407": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 4e-08, + "output_cost_per_token": 1.2e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/NousResearch/Hermes-3-Llama-3.1-405B": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 1e-06, + "output_cost_per_token": 3e-06, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/nvidia/Llama-3.1-Nemotron-Ultra-253B-v1": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 6e-07, + "output_cost_per_token": 1.8e-06, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/nvidia/Llama-3.3-Nemotron-Super-49B-v1": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 4e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/Qwen/Qwen3-235B-A22B": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 6e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/Qwen/Qwen3-32B": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 3e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/Qwen/Qwen3-30B-A3B": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 3e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/Qwen/Qwen3-14B": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 8e-08, + "output_cost_per_token": 2.4e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/Qwen/Qwen3-4B": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 8e-08, + "output_cost_per_token": 2.4e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/Qwen/QwQ-32B": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1.5e-07, + "output_cost_per_token": 4.5e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "supports_reasoning": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/Qwen/Qwen2.5-72B-Instruct": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 1.3e-07, + "output_cost_per_token": 4e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/Qwen/Qwen2.5-32B-Instruct": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 6e-08, + "output_cost_per_token": 2e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/Qwen/Qwen2.5-Coder-7B": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1e-08, + "output_cost_per_token": 3e-08, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/Qwen/Qwen2.5-VL-72B-Instruct": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1.3e-07, + "output_cost_per_token": 4e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "supports_vision": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/Qwen/Qwen2-VL-72B-Instruct": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1.3e-07, + "output_cost_per_token": 4e-07, + "litellm_provider": "nebius", + "mode": "chat", + "supports_function_calling": true, + "supports_vision": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/Qwen/Qwen2-VL-7B-Instruct": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-08, + "output_cost_per_token": 6e-08, + "litellm_provider": "nebius", + "mode": "chat", + "supports_vision": true, + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/BAAI/bge-en-icl": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "input_cost_per_token": 1e-08, + "output_cost_per_token": 0.0, + "litellm_provider": "nebius", + "mode": "embedding", + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/BAAI/bge-multilingual-gemma2": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "input_cost_per_token": 1e-08, + "output_cost_per_token": 0.0, + "litellm_provider": "nebius", + "mode": "embedding", + "source": "https://nebius.com/prices-ai-studio" + }, + "nebius/intfloat/e5-mistral-7b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "input_cost_per_token": 1e-08, + "output_cost_per_token": 0.0, + "litellm_provider": "nebius", + "mode": "embedding", + "source": "https://nebius.com/prices-ai-studio" + }, "nvidia.nemotron-nano-12b-v2": { "input_cost_per_token": 2e-07, "litellm_provider": "bedrock_converse", @@ -26501,6 +22819,16 @@ "supports_system_messages": true, "supports_vision": true }, + "nvidia.nemotron-nano-9b-v2": { + "input_cost_per_token": 6e-08, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 128000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 2.3e-07, + "supports_system_messages": true + }, "nvidia.nemotron-nano-3-30b": { "input_cost_per_token": 6e-08, "litellm_provider": "bedrock_converse", @@ -26514,37 +22842,6 @@ "supports_tool_choice": true, "source": "https://aws.amazon.com/bedrock/pricing/" }, - "nvidia.nemotron-nano-9b-v2": { - "input_cost_per_token": 6e-08, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 128000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 2.3e-07, - "supports_system_messages": true - }, - "nvidia_nim/nvidia/llama-3_2-nv-rerankqa-1b-v2": { - "input_cost_per_query": 0.0, - "input_cost_per_token": 0.0, - "litellm_provider": "nvidia_nim", - "mode": "rerank", - "output_cost_per_token": 0.0 - }, - "nvidia_nim/nvidia/nv-rerankqa-mistral-4b-v3": { - "input_cost_per_query": 0.0, - "input_cost_per_token": 0.0, - "litellm_provider": "nvidia_nim", - "mode": "rerank", - "output_cost_per_token": 0.0 - }, - "nvidia_nim/ranking/nvidia/llama-3.2-nv-rerankqa-1b-v2": { - "input_cost_per_query": 0.0, - "input_cost_per_token": 0.0, - "litellm_provider": "nvidia_nim", - "mode": "rerank", - "output_cost_per_token": 0.0 - }, "o1": { "cache_read_input_token_cost": 7.5e-06, "input_cost_per_token": 1.5e-05, @@ -26997,42 +23294,6 @@ "supports_vision": true, "supports_web_search": true }, - "oci/cohere.command-a-03-2025": { - "input_cost_per_token": 1.56e-06, - "litellm_provider": "oci", - "max_input_tokens": 256000, - "max_output_tokens": 4000, - "max_tokens": 4000, - "mode": "chat", - "output_cost_per_token": 1.56e-06, - "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/", - "supports_function_calling": true, - "supports_response_schema": false - }, - "oci/cohere.command-latest": { - "input_cost_per_token": 1.56e-06, - "litellm_provider": "oci", - "max_input_tokens": 128000, - "max_output_tokens": 4000, - "max_tokens": 4000, - "mode": "chat", - "output_cost_per_token": 1.56e-06, - "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/", - "supports_function_calling": true, - "supports_response_schema": false - }, - "oci/cohere.command-plus-latest": { - "input_cost_per_token": 1.56e-06, - "litellm_provider": "oci", - "max_input_tokens": 128000, - "max_output_tokens": 4000, - "max_tokens": 4000, - "mode": "chat", - "output_cost_per_token": 1.56e-06, - "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/", - "supports_function_calling": true, - "supports_response_schema": false - }, "oci/meta.llama-3.1-405b-instruct": { "input_cost_per_token": 1.068e-05, "litellm_provider": "oci", @@ -27153,6 +23414,42 @@ "supports_function_calling": true, "supports_response_schema": false }, + "oci/cohere.command-latest": { + "input_cost_per_token": 1.56e-06, + "litellm_provider": "oci", + "max_input_tokens": 128000, + "max_output_tokens": 4000, + "max_tokens": 4000, + "mode": "chat", + "output_cost_per_token": 1.56e-06, + "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/", + "supports_function_calling": true, + "supports_response_schema": false + }, + "oci/cohere.command-a-03-2025": { + "input_cost_per_token": 1.56e-06, + "litellm_provider": "oci", + "max_input_tokens": 256000, + "max_output_tokens": 4000, + "max_tokens": 4000, + "mode": "chat", + "output_cost_per_token": 1.56e-06, + "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/", + "supports_function_calling": true, + "supports_response_schema": false + }, + "oci/cohere.command-plus-latest": { + "input_cost_per_token": 1.56e-06, + "litellm_provider": "oci", + "max_input_tokens": 128000, + "max_output_tokens": 4000, + "max_tokens": 4000, + "mode": "chat", + "output_cost_per_token": 1.56e-06, + "source": "https://www.oracle.com/cloud/ai/generative-ai/pricing/", + "supports_function_calling": true, + "supports_response_schema": false + }, "ollama/codegeex4": { "input_cost_per_token": 0.0, "litellm_provider": "ollama", @@ -27495,62 +23792,6 @@ "output_cost_per_token": 2e-07, "supports_system_messages": true }, - "openai/container": { - "code_interpreter_cost_per_session": 0.03, - "litellm_provider": "openai", - "mode": "chat" - }, - "openai/sora-2": { - "litellm_provider": "openai", - "mode": "video_generation", - "output_cost_per_video_per_second": 0.1, - "source": "https://platform.openai.com/docs/api-reference/videos", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "video" - ], - "supported_resolutions": [ - "720x1280", - "1280x720" - ] - }, - "openai/sora-2-pro": { - "litellm_provider": "openai", - "mode": "video_generation", - "output_cost_per_video_per_second": 0.3, - "source": "https://platform.openai.com/docs/api-reference/videos", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "video" - ], - "supported_resolutions": [ - "720x1280", - "1280x720" - ] - }, - "openai/sora-2-pro-high-res": { - "litellm_provider": "openai", - "mode": "video_generation", - "output_cost_per_video_per_second": 0.5, - "source": "https://platform.openai.com/docs/api-reference/videos", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "video" - ], - "supported_resolutions": [ - "1024x1792", - "1792x1024" - ] - }, "openrouter/anthropic/claude-3-haiku": { "input_cost_per_image": 0.0004, "input_cost_per_token": 2.5e-07, @@ -27594,25 +23835,6 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, - "openrouter/anthropic/claude-haiku-4.5": { - "cache_creation_input_token_cost": 1.25e-06, - "cache_read_input_token_cost": 1e-07, - "input_cost_per_token": 1e-06, - "litellm_provider": "openrouter", - "max_input_tokens": 200000, - "max_output_tokens": 200000, - "max_tokens": 200000, - "mode": "chat", - "output_cost_per_token": 5e-06, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, "openrouter/anthropic/claude-opus-4": { "input_cost_per_image": 0.0048, "cache_creation_input_token_cost": 1.875e-05, @@ -27654,6 +23876,54 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, + "openrouter/anthropic/claude-sonnet-4": { + "input_cost_per_image": 0.0048, + "cache_creation_input_token_cost": 3.75e-06, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost": 3e-07, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "litellm_provider": "openrouter", + "max_input_tokens": 1000000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 159 + }, + "openrouter/anthropic/claude-sonnet-4.6": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost": 3e-07, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "litellm_provider": "openrouter", + "max_input_tokens": 1000000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "source": "https://openrouter.ai/anthropic/claude-sonnet-4.6", + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 159 + }, "openrouter/anthropic/claude-opus-4.5": { "cache_creation_input_token_cost": 6.25e-06, "cache_read_input_token_cost": 5e-07, @@ -27692,30 +23962,6 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 346 }, - "openrouter/anthropic/claude-sonnet-4": { - "input_cost_per_image": 0.0048, - "cache_creation_input_token_cost": 3.75e-06, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost": 3e-07, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "litellm_provider": "openrouter", - "max_input_tokens": 1000000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 159 - }, "openrouter/anthropic/claude-sonnet-4.5": { "input_cost_per_image": 0.0048, "cache_creation_input_token_cost": 3.75e-06, @@ -27740,21 +23986,16 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, - "openrouter/anthropic/claude-sonnet-4.6": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost": 3e-07, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, + "openrouter/anthropic/claude-haiku-4.5": { + "cache_creation_input_token_cost": 1.25e-06, + "cache_read_input_token_cost": 1e-07, + "input_cost_per_token": 1e-06, "litellm_provider": "openrouter", - "max_input_tokens": 1000000, - "max_output_tokens": 128000, - "max_tokens": 128000, + "max_input_tokens": 200000, + "max_output_tokens": 200000, + "max_tokens": 200000, "mode": "chat", - "output_cost_per_token": 1.5e-05, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "source": "https://openrouter.ai/anthropic/claude-sonnet-4.6", + "output_cost_per_token": 5e-06, "supports_assistant_prefill": true, "supports_computer_use": true, "supports_function_calling": true, @@ -27762,7 +24003,7 @@ "supports_reasoning": true, "supports_tool_choice": true, "supports_vision": true, - "tool_use_system_prompt_tokens": 159 + "tool_use_system_prompt_tokens": 346 }, "openrouter/bytedance/ui-tars-1.5-7b": { "input_cost_per_token": 1e-07, @@ -27812,36 +24053,6 @@ "supports_reasoning": true, "supports_tool_choice": true }, - "openrouter/deepseek/deepseek-r1": { - "input_cost_per_token": 5.5e-07, - "input_cost_per_token_cache_hit": 1.4e-07, - "litellm_provider": "openrouter", - "max_input_tokens": 65336, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 2.19e-06, - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, - "openrouter/deepseek/deepseek-r1-0528": { - "input_cost_per_token": 5e-07, - "input_cost_per_token_cache_hit": 1.4e-07, - "litellm_provider": "openrouter", - "max_input_tokens": 65336, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 2.15e-06, - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, "openrouter/deepseek/deepseek-v3.2": { "input_cost_per_token": 2.8e-07, "input_cost_per_token_cache_hit": 2.8e-08, @@ -27872,6 +24083,36 @@ "supports_reasoning": false, "supports_tool_choice": true }, + "openrouter/deepseek/deepseek-r1": { + "input_cost_per_token": 5.5e-07, + "input_cost_per_token_cache_hit": 1.4e-07, + "litellm_provider": "openrouter", + "max_input_tokens": 65336, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 2.19e-06, + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, + "openrouter/deepseek/deepseek-r1-0528": { + "input_cost_per_token": 5e-07, + "input_cost_per_token_cache_hit": 1.4e-07, + "litellm_provider": "openrouter", + "max_input_tokens": 65336, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 2.15e-06, + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, "openrouter/google/gemini-2.0-flash-001": { "deprecation_date": "2026-06-01", "input_cost_per_audio_token": 7e-07, @@ -27939,53 +24180,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "openrouter/google/gemini-3-flash-preview": { - "cache_read_input_token_cost": 5e-08, - "input_cost_per_audio_token": 1e-06, - "input_cost_per_token": 5e-07, - "litellm_provider": "openrouter", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_reasoning_token": 3e-06, - "output_cost_per_token": 3e-06, - "rpm": 2000, - "source": "https://ai.google.dev/pricing/gemini-3", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_output": false, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_url_context": true, - "supports_vision": true, - "supports_web_search": true, - "tpm": 800000 - }, "openrouter/google/gemini-3-pro-preview": { "cache_read_input_token_cost": 2e-07, "cache_read_input_token_cost_above_200k_tokens": 4e-07, @@ -28033,6 +24227,53 @@ "supports_vision": true, "supports_web_search": true }, + "openrouter/google/gemini-3-flash-preview": { + "cache_read_input_token_cost": 5e-08, + "input_cost_per_audio_token": 1e-06, + "input_cost_per_token": 5e-07, + "litellm_provider": "openrouter", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_reasoning_token": 3e-06, + "output_cost_per_token": 3e-06, + "rpm": 2000, + "source": "https://ai.google.dev/pricing/gemini-3", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true, + "tpm": 800000 + }, "openrouter/google/gemini-3.1-pro-preview": { "cache_read_input_token_cost": 2e-07, "cache_read_input_token_cost_above_200k_tokens": 4e-07, @@ -28103,40 +24344,6 @@ "supports_reasoning": true, "supports_tool_choice": true }, - "openrouter/minimax/minimax-m2.1": { - "input_cost_per_token": 2.7e-07, - "output_cost_per_token": 1.2e-06, - "cache_creation_input_token_cost": 0.0, - "cache_read_input_token_cost": 0.0, - "litellm_provider": "openrouter", - "max_input_tokens": 204000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_reasoning": true, - "supports_vision": true, - "supports_prompt_caching": false, - "supports_computer_use": false - }, - "openrouter/minimax/minimax-m2.5": { - "input_cost_per_token": 3e-07, - "output_cost_per_token": 1.1e-06, - "cache_read_input_token_cost": 1.5e-07, - "litellm_provider": "openrouter", - "max_input_tokens": 196608, - "max_output_tokens": 65536, - "max_tokens": 65536, - "mode": "chat", - "source": "https://openrouter.ai/minimax/minimax-m2.5", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_reasoning": true, - "supports_vision": false, - "supports_prompt_caching": true, - "supports_computer_use": false - }, "openrouter/mistralai/devstral-2512": { "input_cost_per_image": 0, "input_cost_per_token": 1.5e-07, @@ -28151,20 +24358,6 @@ "supports_tool_choice": true, "supports_vision": false }, - "openrouter/mistralai/ministral-14b-2512": { - "input_cost_per_image": 0, - "input_cost_per_token": 2e-07, - "litellm_provider": "openrouter", - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 2e-07, - "supports_function_calling": true, - "supports_prompt_caching": false, - "supports_tool_choice": true, - "supports_vision": true - }, "openrouter/mistralai/ministral-3b-2512": { "input_cost_per_image": 0, "input_cost_per_token": 1e-07, @@ -28193,6 +24386,34 @@ "supports_tool_choice": true, "supports_vision": true }, + "openrouter/mistralai/ministral-14b-2512": { + "input_cost_per_image": 0, + "input_cost_per_token": 2e-07, + "litellm_provider": "openrouter", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 2e-07, + "supports_function_calling": true, + "supports_prompt_caching": false, + "supports_tool_choice": true, + "supports_vision": true + }, + "openrouter/mistralai/mistral-large-2512": { + "input_cost_per_image": 0, + "input_cost_per_token": 5e-07, + "litellm_provider": "openrouter", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 1.5e-06, + "supports_function_calling": true, + "supports_prompt_caching": false, + "supports_tool_choice": true, + "supports_vision": true + }, "openrouter/mistralai/mistral-7b-instruct": { "input_cost_per_token": 1.3e-07, "litellm_provider": "openrouter", @@ -28209,20 +24430,6 @@ "output_cost_per_token": 2.4e-05, "supports_tool_choice": true }, - "openrouter/mistralai/mistral-large-2512": { - "input_cost_per_image": 0, - "input_cost_per_token": 5e-07, - "litellm_provider": "openrouter", - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 1.5e-06, - "supports_function_calling": true, - "supports_prompt_caching": false, - "supports_tool_choice": true, - "supports_vision": true - }, "openrouter/mistralai/mistral-small-3.1-24b-instruct": { "input_cost_per_token": 1e-07, "litellm_provider": "openrouter", @@ -28363,25 +24570,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "openrouter/openai/gpt-5": { - "cache_read_input_token_cost": 1.25e-07, - "input_cost_per_token": 1.25e-06, - "litellm_provider": "openrouter", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1e-05, - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_reasoning": true, - "supports_tool_choice": true - }, "openrouter/openai/gpt-5-chat": { "cache_read_input_token_cost": 1.25e-07, "input_cost_per_token": 1.25e-06, @@ -28420,6 +24608,44 @@ "supports_reasoning": true, "supports_tool_choice": true }, + "openrouter/openai/gpt-5.2-codex": { + "cache_read_input_token_cost": 1.75e-07, + "input_cost_per_token": 1.75e-06, + "litellm_provider": "openrouter", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1.4e-05, + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_reasoning": true, + "supports_tool_choice": true + }, + "openrouter/openai/gpt-5": { + "cache_read_input_token_cost": 1.25e-07, + "input_cost_per_token": 1.25e-06, + "litellm_provider": "openrouter", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1e-05, + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text" + ], + "supports_reasoning": true, + "supports_tool_choice": true + }, "openrouter/openai/gpt-5-mini": { "cache_read_input_token_cost": 2.5e-08, "input_cost_per_token": 2.5e-07, @@ -28512,25 +24738,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "openrouter/openai/gpt-5.2-codex": { - "cache_read_input_token_cost": 1.75e-07, - "input_cost_per_token": 1.75e-06, - "litellm_provider": "openrouter", - "max_input_tokens": 272000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 1.4e-05, - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "text" - ], - "supports_reasoning": true, - "supports_tool_choice": true - }, "openrouter/openai/gpt-5.2-pro": { "input_cost_per_image": 0, "input_cost_per_token": 2.1e-05, @@ -28621,42 +24828,6 @@ "supports_tool_choice": true, "supports_vision": false }, - "openrouter/openrouter/auto": { - "input_cost_per_token": 0, - "output_cost_per_token": 0, - "litellm_provider": "openrouter", - "max_input_tokens": 2000000, - "max_tokens": 2000000, - "mode": "chat", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_vision": true, - "supports_audio_input": true, - "supports_video_input": true - }, - "openrouter/openrouter/bodybuilder": { - "input_cost_per_token": 0, - "output_cost_per_token": 0, - "litellm_provider": "openrouter", - "max_input_tokens": 128000, - "max_tokens": 128000, - "mode": "chat" - }, - "openrouter/openrouter/free": { - "input_cost_per_token": 0, - "output_cost_per_token": 0, - "litellm_provider": "openrouter", - "max_input_tokens": 200000, - "max_tokens": 200000, - "mode": "chat", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_vision": true - }, "openrouter/qwen/qwen-2.5-coder-32b-instruct": { "input_cost_per_token": 1.8e-07, "litellm_provider": "openrouter", @@ -28678,31 +24849,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "openrouter/qwen/qwen3-235b-a22b-2507": { - "input_cost_per_token": 7.1e-08, - "litellm_provider": "openrouter", - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 1e-07, - "source": "https://openrouter.ai/qwen/qwen3-235b-a22b-2507", - "supports_function_calling": true, - "supports_tool_choice": true - }, - "openrouter/qwen/qwen3-235b-a22b-thinking-2507": { - "input_cost_per_token": 1.1e-07, - "litellm_provider": "openrouter", - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "max_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 6e-07, - "source": "https://openrouter.ai/qwen/qwen3-235b-a22b-thinking-2507", - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, "openrouter/qwen/qwen3-coder": { "input_cost_per_token": 2.2e-07, "litellm_provider": "openrouter", @@ -28728,15 +24874,40 @@ "supports_reasoning": true, "supports_tool_choice": true }, - "openrouter/qwen/qwen3.5-122b-a10b": { - "input_cost_per_token": 4e-07, + "openrouter/qwen/qwen3-235b-a22b-2507": { + "input_cost_per_token": 7.1e-08, + "litellm_provider": "openrouter", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 1e-07, + "source": "https://openrouter.ai/qwen/qwen3-235b-a22b-2507", + "supports_function_calling": true, + "supports_tool_choice": true + }, + "openrouter/qwen/qwen3-235b-a22b-thinking-2507": { + "input_cost_per_token": 1.1e-07, + "litellm_provider": "openrouter", + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 6e-07, + "source": "https://openrouter.ai/qwen/qwen3-235b-a22b-thinking-2507", + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, + "openrouter/qwen/qwen3.5-35b-a3b": { + "input_cost_per_token": 2.5e-07, "litellm_provider": "openrouter", "max_input_tokens": 262144, "max_output_tokens": 65536, "max_tokens": 65536, "mode": "chat", "output_cost_per_token": 2e-06, - "source": "https://openrouter.ai/qwen/qwen3.5-122b-a10b", + "source": "https://openrouter.ai/qwen/qwen3.5-35b-a3b", "supports_function_calling": true, "supports_reasoning": true, "supports_tool_choice": true, @@ -28756,29 +24927,15 @@ "supports_tool_choice": true, "supports_vision": true }, - "openrouter/qwen/qwen3.5-35b-a3b": { - "input_cost_per_token": 2.5e-07, + "openrouter/qwen/qwen3.5-122b-a10b": { + "input_cost_per_token": 4e-07, "litellm_provider": "openrouter", "max_input_tokens": 262144, "max_output_tokens": 65536, "max_tokens": 65536, "mode": "chat", "output_cost_per_token": 2e-06, - "source": "https://openrouter.ai/qwen/qwen3.5-35b-a3b", - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true, - "supports_vision": true - }, - "openrouter/qwen/qwen3.5-397b-a17b": { - "input_cost_per_token": 6e-07, - "litellm_provider": "openrouter", - "max_input_tokens": 262144, - "max_output_tokens": 65536, - "max_tokens": 65536, - "mode": "chat", - "output_cost_per_token": 3.6e-06, - "source": "https://openrouter.ai/qwen/qwen3.5-397b-a17b", + "source": "https://openrouter.ai/qwen/qwen3.5-122b-a10b", "supports_function_calling": true, "supports_reasoning": true, "supports_tool_choice": true, @@ -28814,6 +24971,20 @@ "supports_tool_choice": true, "supports_vision": true }, + "openrouter/qwen/qwen3.5-397b-a17b": { + "input_cost_per_token": 6e-07, + "litellm_provider": "openrouter", + "max_input_tokens": 262144, + "max_output_tokens": 65536, + "max_tokens": 65536, + "mode": "chat", + "output_cost_per_token": 3.6e-06, + "source": "https://openrouter.ai/qwen/qwen3.5-397b-a17b", + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "supports_vision": true + }, "openrouter/switchpoint/router": { "input_cost_per_token": 8.5e-07, "litellm_provider": "openrouter", @@ -28847,22 +25018,6 @@ "supports_tool_choice": true, "supports_web_search": true }, - "openrouter/xiaomi/mimo-v2-flash": { - "input_cost_per_token": 9e-08, - "output_cost_per_token": 2.9e-07, - "cache_creation_input_token_cost": 0.0, - "cache_read_input_token_cost": 0.0, - "litellm_provider": "openrouter", - "max_input_tokens": 262144, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_reasoning": true, - "supports_vision": false, - "supports_prompt_caching": false - }, "openrouter/z-ai/glm-4.6": { "input_cost_per_token": 4e-07, "litellm_provider": "openrouter", @@ -28891,6 +25046,22 @@ "supports_reasoning": true, "supports_tool_choice": true }, + "openrouter/xiaomi/mimo-v2-flash": { + "input_cost_per_token": 9e-08, + "output_cost_per_token": 2.9e-07, + "cache_creation_input_token_cost": 0.0, + "cache_read_input_token_cost": 0.0, + "litellm_provider": "openrouter", + "max_input_tokens": 262144, + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_reasoning": true, + "supports_vision": false, + "supports_prompt_caching": false + }, "openrouter/z-ai/glm-4.7": { "input_cost_per_token": 4e-07, "output_cost_per_token": 1.5e-06, @@ -28937,6 +25108,76 @@ "supports_reasoning": true, "supports_tool_choice": true }, + "openrouter/minimax/minimax-m2.1": { + "input_cost_per_token": 2.7e-07, + "output_cost_per_token": 1.2e-06, + "cache_creation_input_token_cost": 0.0, + "cache_read_input_token_cost": 0.0, + "litellm_provider": "openrouter", + "max_input_tokens": 204000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_reasoning": true, + "supports_vision": true, + "supports_prompt_caching": false, + "supports_computer_use": false + }, + "openrouter/minimax/minimax-m2.5": { + "input_cost_per_token": 3e-07, + "output_cost_per_token": 1.1e-06, + "cache_read_input_token_cost": 1.5e-07, + "litellm_provider": "openrouter", + "max_input_tokens": 196608, + "max_output_tokens": 65536, + "max_tokens": 65536, + "mode": "chat", + "source": "https://openrouter.ai/minimax/minimax-m2.5", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_reasoning": true, + "supports_vision": false, + "supports_prompt_caching": true, + "supports_computer_use": false + }, + "openrouter/openrouter/auto": { + "input_cost_per_token": 0, + "output_cost_per_token": 0, + "litellm_provider": "openrouter", + "max_input_tokens": 2000000, + "max_tokens": 2000000, + "mode": "chat", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_vision": true, + "supports_audio_input": true, + "supports_video_input": true + }, + "openrouter/openrouter/free": { + "input_cost_per_token": 0, + "output_cost_per_token": 0, + "litellm_provider": "openrouter", + "max_input_tokens": 200000, + "max_tokens": 200000, + "mode": "chat", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_vision": true + }, + "openrouter/openrouter/bodybuilder": { + "input_cost_per_token": 0, + "output_cost_per_token": 0, + "litellm_provider": "openrouter", + "max_input_tokens": 128000, + "max_tokens": 128000, + "mode": "chat" + }, "ovhcloud/DeepSeek-R1-Distill-Llama-70B": { "input_cost_per_token": 6.7e-07, "litellm_provider": "ovhcloud", @@ -29209,34 +25450,6 @@ "litellm_provider": "parallel_ai", "mode": "search" }, - "perplexity/anthropic/claude-haiku-4-5": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": false, - "supports_function_calling": true - }, - "perplexity/anthropic/claude-opus-4-5": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": false, - "supports_function_calling": true - }, - "perplexity/anthropic/claude-opus-4-6": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": false, - "supports_function_calling": true - }, - "perplexity/anthropic/claude-sonnet-4-5": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": false, - "supports_function_calling": true - }, "perplexity/codellama-34b-instruct": { "input_cost_per_token": 3.5e-07, "litellm_provider": "perplexity", @@ -29255,34 +25468,6 @@ "mode": "chat", "output_cost_per_token": 2.8e-06 }, - "perplexity/google/gemini-2.5-flash": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": false, - "supports_function_calling": true - }, - "perplexity/google/gemini-2.5-pro": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": false, - "supports_function_calling": true - }, - "perplexity/google/gemini-3-flash-preview": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": false, - "supports_function_calling": true - }, - "perplexity/google/gemini-3-pro-preview": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": false, - "supports_function_calling": true - }, "perplexity/llama-2-70b-chat": { "input_cost_per_token": 7e-07, "litellm_provider": "perplexity", @@ -29328,34 +25513,6 @@ "mode": "chat", "output_cost_per_token": 2.8e-07 }, - "perplexity/openai/gpt-5-mini": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": false, - "supports_function_calling": true - }, - "perplexity/openai/gpt-5.1": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": false, - "supports_function_calling": true - }, - "perplexity/openai/gpt-5.2": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": true, - "supports_function_calling": true - }, - "perplexity/perplexity/sonar": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": false, - "supports_function_calling": true - }, "perplexity/pplx-70b-chat": { "input_cost_per_token": 7e-07, "litellm_provider": "perplexity", @@ -29394,59 +25551,6 @@ "mode": "chat", "output_cost_per_token": 2.8e-07 }, - "perplexity/pplx-embed-v1-0.6b": { - "input_cost_per_token": 4e-09, - "litellm_provider": "perplexity", - "max_input_tokens": 32768, - "max_tokens": 32768, - "mode": "embedding", - "output_cost_per_token": 0.0, - "output_vector_size": 1024, - "source": "https://docs.perplexity.ai/docs/embeddings/quickstart" - }, - "perplexity/pplx-embed-v1-4b": { - "input_cost_per_token": 3e-08, - "litellm_provider": "perplexity", - "max_input_tokens": 32768, - "max_tokens": 32768, - "mode": "embedding", - "output_cost_per_token": 0.0, - "output_vector_size": 2560, - "source": "https://docs.perplexity.ai/docs/embeddings/quickstart" - }, - "perplexity/preset/advanced-deep-research": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_preset": true, - "supports_function_calling": true - }, - "perplexity/preset/deep-research": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_preset": true, - "supports_function_calling": true - }, - "perplexity/preset/fast-search": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_preset": true, - "supports_function_calling": true - }, - "perplexity/preset/pro-search": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_preset": true, - "supports_function_calling": true - }, - "perplexity/search": { - "input_cost_per_query": 0.005, - "litellm_provider": "perplexity", - "mode": "search" - }, "perplexity/sonar": { "input_cost_per_token": 1e-06, "litellm_provider": "perplexity", @@ -29561,14 +25665,31 @@ "mode": "chat", "output_cost_per_token": 2.8e-07 }, - "perplexity/xai/grok-4-1-fast-non-reasoning": { - "litellm_provider": "perplexity", - "mode": "responses", - "supports_web_search": true, - "supports_reasoning": false, - "supports_function_calling": true + "publicai/swiss-ai/apertus-8b-instruct": { + "input_cost_per_token": 0.0, + "litellm_provider": "publicai", + "max_input_tokens": 8192, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_token": 0.0, + "source": "https://platform.publicai.co/docs", + "supports_function_calling": false, + "supports_tool_choice": false }, - "publicai/BSC-LT/ALIA-40b-instruct_Q8_0": { + "publicai/swiss-ai/apertus-70b-instruct": { + "input_cost_per_token": 0.0, + "litellm_provider": "publicai", + "max_input_tokens": 8192, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_token": 0.0, + "source": "https://platform.publicai.co/docs", + "supports_function_calling": false, + "supports_tool_choice": false + }, + "publicai/aisingapore/Gemma-SEA-LION-v4-27B-IT": { "input_cost_per_token": 0.0, "litellm_provider": "publicai", "max_input_tokens": 8192, @@ -29592,7 +25713,7 @@ "supports_function_calling": true, "supports_tool_choice": true }, - "publicai/aisingapore/Gemma-SEA-LION-v4-27B-IT": { + "publicai/BSC-LT/ALIA-40b-instruct_Q8_0": { "input_cost_per_token": 0.0, "litellm_provider": "publicai", "max_input_tokens": 8192, @@ -29604,7 +25725,7 @@ "supports_function_calling": true, "supports_tool_choice": true }, - "publicai/aisingapore/Qwen-SEA-LION-v4-32B-IT": { + "publicai/allenai/Olmo-3-7B-Instruct": { "input_cost_per_token": 0.0, "litellm_provider": "publicai", "max_input_tokens": 32768, @@ -29616,20 +25737,146 @@ "supports_function_calling": true, "supports_tool_choice": true }, - "publicai/allenai/Olmo-3-32B-Think": { - "input_cost_per_token": 0.0, - "litellm_provider": "publicai", - "max_input_tokens": 32768, - "max_output_tokens": 4096, - "max_tokens": 4096, - "mode": "chat", - "output_cost_per_token": 0.0, - "source": "https://platform.publicai.co/docs", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_reasoning": true + "perplexity/preset/fast-search": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_preset": true, + "supports_function_calling": true }, - "publicai/allenai/Olmo-3-7B-Instruct": { + "perplexity/preset/pro-search": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_preset": true, + "supports_function_calling": true + }, + "perplexity/preset/deep-research": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_preset": true, + "supports_function_calling": true + }, + "perplexity/preset/advanced-deep-research": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_preset": true, + "supports_function_calling": true + }, + "perplexity/openai/gpt-5.2": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": true, + "supports_function_calling": true + }, + "perplexity/openai/gpt-5.1": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": false, + "supports_function_calling": true + }, + "perplexity/openai/gpt-5-mini": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": false, + "supports_function_calling": true + }, + "perplexity/anthropic/claude-opus-4-6": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": false, + "supports_function_calling": true + }, + "perplexity/anthropic/claude-opus-4-5": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": false, + "supports_function_calling": true + }, + "perplexity/anthropic/claude-sonnet-4-5": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": false, + "supports_function_calling": true + }, + "perplexity/anthropic/claude-haiku-4-5": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": false, + "supports_function_calling": true + }, + "perplexity/google/gemini-3-pro-preview": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": false, + "supports_function_calling": true + }, + "perplexity/google/gemini-3-flash-preview": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": false, + "supports_function_calling": true + }, + "perplexity/google/gemini-2.5-pro": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": false, + "supports_function_calling": true + }, + "perplexity/google/gemini-2.5-flash": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": false, + "supports_function_calling": true + }, + "perplexity/xai/grok-4-1-fast-non-reasoning": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": false, + "supports_function_calling": true + }, + "perplexity/perplexity/sonar": { + "litellm_provider": "perplexity", + "mode": "responses", + "supports_web_search": true, + "supports_reasoning": false, + "supports_function_calling": true + }, + "perplexity/pplx-embed-v1-0.6b": { + "input_cost_per_token": 4e-09, + "litellm_provider": "perplexity", + "max_input_tokens": 32768, + "max_tokens": 32768, + "mode": "embedding", + "output_cost_per_token": 0.0, + "output_vector_size": 1024, + "source": "https://docs.perplexity.ai/docs/embeddings/quickstart" + }, + "perplexity/pplx-embed-v1-4b": { + "input_cost_per_token": 3e-08, + "litellm_provider": "perplexity", + "max_input_tokens": 32768, + "max_tokens": 32768, + "mode": "embedding", + "output_cost_per_token": 0.0, + "output_vector_size": 2560, + "source": "https://docs.perplexity.ai/docs/embeddings/quickstart" + }, + "publicai/aisingapore/Qwen-SEA-LION-v4-32B-IT": { "input_cost_per_token": 0.0, "litellm_provider": "publicai", "max_input_tokens": 32768, @@ -29654,29 +25901,30 @@ "supports_tool_choice": true, "supports_reasoning": true }, - "publicai/swiss-ai/apertus-70b-instruct": { + "publicai/allenai/Olmo-3-32B-Think": { "input_cost_per_token": 0.0, "litellm_provider": "publicai", - "max_input_tokens": 8192, + "max_input_tokens": 32768, "max_output_tokens": 4096, "max_tokens": 4096, "mode": "chat", "output_cost_per_token": 0.0, "source": "https://platform.publicai.co/docs", - "supports_function_calling": false, - "supports_tool_choice": false + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_reasoning": true }, - "publicai/swiss-ai/apertus-8b-instruct": { - "input_cost_per_token": 0.0, - "litellm_provider": "publicai", - "max_input_tokens": 8192, - "max_output_tokens": 4096, - "max_tokens": 4096, + "qwen.qwen3-coder-480b-a35b-v1:0": { + "input_cost_per_token": 2.2e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 262000, + "max_output_tokens": 65536, + "max_tokens": 65536, "mode": "chat", - "output_cost_per_token": 0.0, - "source": "https://platform.publicai.co/docs", - "supports_function_calling": false, - "supports_tool_choice": false + "output_cost_per_token": 1.8e-06, + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true }, "qwen.qwen3-235b-a22b-2507-v1:0": { "input_cost_per_token": 2.2e-07, @@ -29690,18 +25938,6 @@ "supports_reasoning": true, "supports_tool_choice": true }, - "qwen.qwen3-32b-v1:0": { - "input_cost_per_token": 1.5e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 131072, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 6e-07, - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, "qwen.qwen3-coder-30b-a3b-v1:0": { "input_cost_per_token": 1.5e-07, "litellm_provider": "bedrock_converse", @@ -29714,31 +25950,18 @@ "supports_reasoning": true, "supports_tool_choice": true }, - "qwen.qwen3-coder-480b-a35b-v1:0": { - "input_cost_per_token": 2.2e-07, + "qwen.qwen3-32b-v1:0": { + "input_cost_per_token": 1.5e-07, "litellm_provider": "bedrock_converse", - "max_input_tokens": 262000, - "max_output_tokens": 65536, - "max_tokens": 65536, + "max_input_tokens": 131072, + "max_output_tokens": 16384, + "max_tokens": 16384, "mode": "chat", - "output_cost_per_token": 1.8e-06, + "output_cost_per_token": 6e-07, "supports_function_calling": true, "supports_reasoning": true, "supports_tool_choice": true }, - "qwen.qwen3-coder-next": { - "input_cost_per_token": 5e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 262144, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1.2e-06, - "supports_function_calling": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "source": "https://aws.amazon.com/bedrock/pricing/" - }, "qwen.qwen3-next-80b-a3b": { "input_cost_per_token": 1.5e-07, "litellm_provider": "bedrock_converse", @@ -29762,6 +25985,19 @@ "supports_system_messages": true, "supports_vision": true }, + "qwen.qwen3-coder-next": { + "input_cost_per_token": 5e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 262144, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 1.2e-06, + "supports_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "source": "https://aws.amazon.com/bedrock/pricing/" + }, "recraft/recraftv2": { "litellm_provider": "recraft", "mode": "image_generation", @@ -29780,151 +26016,6 @@ "/v1/images/generations" ] }, - "replicate/anthropic/claude-3.5-haiku": { - "input_cost_per_token": 1e-06, - "output_cost_per_token": 5e-06, - "litellm_provider": "replicate", - "mode": "chat", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_response_schema": true, - "supports_prompt_caching": true - }, - "replicate/anthropic/claude-3.5-sonnet": { - "input_cost_per_token": 3.75e-06, - "output_cost_per_token": 1.875e-05, - "litellm_provider": "replicate", - "mode": "chat", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_response_schema": true, - "supports_prompt_caching": true - }, - "replicate/anthropic/claude-3.7-sonnet": { - "input_cost_per_token": 3e-06, - "output_cost_per_token": 1.5e-05, - "litellm_provider": "replicate", - "mode": "chat", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_response_schema": true, - "supports_prompt_caching": true - }, - "replicate/anthropic/claude-4-sonnet": { - "input_cost_per_token": 3e-06, - "output_cost_per_token": 1.5e-05, - "litellm_provider": "replicate", - "mode": "chat", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_response_schema": true, - "supports_prompt_caching": true - }, - "replicate/anthropic/claude-4.5-haiku": { - "input_cost_per_token": 1e-06, - "output_cost_per_token": 5e-06, - "litellm_provider": "replicate", - "mode": "chat", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_response_schema": true, - "supports_prompt_caching": true - }, - "replicate/anthropic/claude-4.5-sonnet": { - "input_cost_per_token": 3e-06, - "output_cost_per_token": 1.5e-05, - "litellm_provider": "replicate", - "mode": "chat", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_response_schema": true, - "supports_prompt_caching": true - }, - "replicate/deepseek-ai/deepseek-r1": { - "input_cost_per_token": 3.75e-06, - "output_cost_per_token": 1e-05, - "output_cost_per_reasoning_token": 1e-05, - "litellm_provider": "replicate", - "mode": "chat", - "max_input_tokens": 65536, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_reasoning": true, - "supports_system_messages": true - }, - "replicate/deepseek-ai/deepseek-v3": { - "input_cost_per_token": 1.45e-06, - "output_cost_per_token": 1.45e-06, - "litellm_provider": "replicate", - "mode": "chat", - "max_input_tokens": 65536, - "max_output_tokens": 8192, - "max_tokens": 8192, - "supports_function_calling": true, - "supports_system_messages": true - }, - "replicate/deepseek-ai/deepseek-v3.1": { - "input_cost_per_token": 6.72e-07, - "output_cost_per_token": 2.016e-06, - "litellm_provider": "replicate", - "mode": "chat", - "max_input_tokens": 163840, - "max_output_tokens": 163840, - "max_tokens": 163840, - "supports_function_calling": true, - "supports_reasoning": true, - "supports_system_messages": true - }, - "replicate/google/gemini-2.5-flash": { - "input_cost_per_token": 2.5e-06, - "output_cost_per_token": 2.5e-06, - "litellm_provider": "replicate", - "mode": "chat", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_response_schema": true - }, - "replicate/google/gemini-3-pro": { - "input_cost_per_token": 2e-06, - "output_cost_per_token": 1.2e-05, - "litellm_provider": "replicate", - "mode": "chat", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_vision": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_response_schema": true - }, - "replicate/ibm-granite/granite-3.3-8b-instruct": { - "input_cost_per_token": 3e-08, - "output_cost_per_token": 2.5e-07, - "litellm_provider": "replicate", - "mode": "chat", - "supports_function_calling": true, - "supports_system_messages": true - }, "replicate/meta/llama-2-13b": { "input_cost_per_token": 1e-07, "litellm_provider": "replicate", @@ -30055,9 +26146,9 @@ "output_cost_per_token": 1e-06, "supports_tool_choice": true }, - "replicate/openai/gpt-4.1": { - "input_cost_per_token": 2e-06, - "output_cost_per_token": 8e-06, + "replicate/openai/gpt-5": { + "input_cost_per_token": 1.25e-06, + "output_cost_per_token": 1e-05, "litellm_provider": "replicate", "mode": "chat", "supports_function_calling": true, @@ -30067,9 +26158,17 @@ "supports_tool_choice": true, "supports_response_schema": true }, - "replicate/openai/gpt-4.1-mini": { - "input_cost_per_token": 4e-07, - "output_cost_per_token": 1.6e-06, + "replicateopenai/gpt-oss-20b": { + "input_cost_per_token": 9e-08, + "output_cost_per_token": 3.6e-07, + "litellm_provider": "replicate", + "mode": "chat", + "supports_function_calling": true, + "supports_system_messages": true + }, + "replicate/anthropic/claude-4.5-haiku": { + "input_cost_per_token": 1e-06, + "output_cost_per_token": 5e-06, "litellm_provider": "replicate", "mode": "chat", "supports_function_calling": true, @@ -30077,11 +26176,12 @@ "supports_vision": true, "supports_system_messages": true, "supports_tool_choice": true, - "supports_response_schema": true + "supports_response_schema": true, + "supports_prompt_caching": true }, - "replicate/openai/gpt-4.1-nano": { - "input_cost_per_token": 1e-07, - "output_cost_per_token": 4e-07, + "replicate/ibm-granite/granite-3.3-8b-instruct": { + "input_cost_per_token": 3e-08, + "output_cost_per_token": 2.5e-07, "litellm_provider": "replicate", "mode": "chat", "supports_function_calling": true, @@ -30101,6 +26201,33 @@ "supports_audio_input": true, "supports_audio_output": true }, + "replicate/openai/o4-mini": { + "input_cost_per_token": 1e-06, + "output_cost_per_token": 4e-06, + "output_cost_per_reasoning_token": 4e-06, + "litellm_provider": "replicate", + "mode": "chat", + "supports_reasoning": true, + "supports_system_messages": true + }, + "replicate/openai/o1-mini": { + "input_cost_per_token": 1.1e-06, + "output_cost_per_token": 4.4e-06, + "output_cost_per_reasoning_token": 4.4e-06, + "litellm_provider": "replicate", + "mode": "chat", + "supports_reasoning": true, + "supports_system_messages": true + }, + "replicate/openai/o1": { + "input_cost_per_token": 1.5e-05, + "output_cost_per_token": 6e-05, + "output_cost_per_reasoning_token": 6e-05, + "litellm_provider": "replicate", + "mode": "chat", + "supports_reasoning": true, + "supports_system_messages": true + }, "replicate/openai/gpt-4o-mini": { "input_cost_per_token": 1.5e-07, "output_cost_per_token": 6e-07, @@ -30113,9 +26240,80 @@ "supports_tool_choice": true, "supports_response_schema": true }, - "replicate/openai/gpt-5": { - "input_cost_per_token": 1.25e-06, - "output_cost_per_token": 1e-05, + "replicate/qwen/qwen3-235b-a22b-instruct-2507": { + "input_cost_per_token": 2.64e-07, + "output_cost_per_token": 1.06e-06, + "litellm_provider": "replicate", + "mode": "chat", + "supports_function_calling": true, + "supports_system_messages": true + }, + "replicate/anthropic/claude-4-sonnet": { + "input_cost_per_token": 3e-06, + "output_cost_per_token": 1.5e-05, + "litellm_provider": "replicate", + "mode": "chat", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_response_schema": true, + "supports_prompt_caching": true + }, + "replicate/deepseek-ai/deepseek-v3": { + "input_cost_per_token": 1.45e-06, + "output_cost_per_token": 1.45e-06, + "litellm_provider": "replicate", + "mode": "chat", + "max_input_tokens": 65536, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_function_calling": true, + "supports_system_messages": true + }, + "replicate/anthropic/claude-3.7-sonnet": { + "input_cost_per_token": 3e-06, + "output_cost_per_token": 1.5e-05, + "litellm_provider": "replicate", + "mode": "chat", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_response_schema": true, + "supports_prompt_caching": true + }, + "replicate/anthropic/claude-3.5-haiku": { + "input_cost_per_token": 1e-06, + "output_cost_per_token": 5e-06, + "litellm_provider": "replicate", + "mode": "chat", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_response_schema": true, + "supports_prompt_caching": true + }, + "replicate/anthropic/claude-3.5-sonnet": { + "input_cost_per_token": 3.75e-06, + "output_cost_per_token": 1.875e-05, + "litellm_provider": "replicate", + "mode": "chat", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_response_schema": true, + "supports_prompt_caching": true + }, + "replicate/google/gemini-3-pro": { + "input_cost_per_token": 2e-06, + "output_cost_per_token": 1.2e-05, "litellm_provider": "replicate", "mode": "chat", "supports_function_calling": true, @@ -30125,9 +26323,42 @@ "supports_tool_choice": true, "supports_response_schema": true }, - "replicate/openai/gpt-5-mini": { - "input_cost_per_token": 2.5e-07, - "output_cost_per_token": 2e-06, + "replicate/anthropic/claude-4.5-sonnet": { + "input_cost_per_token": 3e-06, + "output_cost_per_token": 1.5e-05, + "litellm_provider": "replicate", + "mode": "chat", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_response_schema": true, + "supports_prompt_caching": true + }, + "replicate/openai/gpt-4.1": { + "input_cost_per_token": 2e-06, + "output_cost_per_token": 8e-06, + "litellm_provider": "replicate", + "mode": "chat", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_response_schema": true + }, + "replicate/openai/gpt-4.1-nano": { + "input_cost_per_token": 1e-07, + "output_cost_per_token": 4e-07, + "litellm_provider": "replicate", + "mode": "chat", + "supports_function_calling": true, + "supports_system_messages": true + }, + "replicate/openai/gpt-4.1-mini": { + "input_cost_per_token": 4e-07, + "output_cost_per_token": 1.6e-06, "litellm_provider": "replicate", "mode": "chat", "supports_function_calling": true, @@ -30145,6 +26376,30 @@ "supports_function_calling": true, "supports_system_messages": true }, + "replicate/openai/gpt-5-mini": { + "input_cost_per_token": 2.5e-07, + "output_cost_per_token": 2e-06, + "litellm_provider": "replicate", + "mode": "chat", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_response_schema": true + }, + "replicate/google/gemini-2.5-flash": { + "input_cost_per_token": 2.5e-06, + "output_cost_per_token": 2.5e-06, + "litellm_provider": "replicate", + "mode": "chat", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_response_schema": true + }, "replicate/openai/gpt-oss-120b": { "input_cost_per_token": 1.8e-07, "output_cost_per_token": 7.2e-07, @@ -30153,39 +26408,16 @@ "supports_function_calling": true, "supports_system_messages": true }, - "replicate/openai/o1": { - "input_cost_per_token": 1.5e-05, - "output_cost_per_token": 6e-05, - "output_cost_per_reasoning_token": 6e-05, - "litellm_provider": "replicate", - "mode": "chat", - "supports_reasoning": true, - "supports_system_messages": true - }, - "replicate/openai/o1-mini": { - "input_cost_per_token": 1.1e-06, - "output_cost_per_token": 4.4e-06, - "output_cost_per_reasoning_token": 4.4e-06, - "litellm_provider": "replicate", - "mode": "chat", - "supports_reasoning": true, - "supports_system_messages": true - }, - "replicate/openai/o4-mini": { - "input_cost_per_token": 1e-06, - "output_cost_per_token": 4e-06, - "output_cost_per_reasoning_token": 4e-06, - "litellm_provider": "replicate", - "mode": "chat", - "supports_reasoning": true, - "supports_system_messages": true - }, - "replicate/qwen/qwen3-235b-a22b-instruct-2507": { - "input_cost_per_token": 2.64e-07, - "output_cost_per_token": 1.06e-06, + "replicate/deepseek-ai/deepseek-v3.1": { + "input_cost_per_token": 6.72e-07, + "output_cost_per_token": 2.016e-06, "litellm_provider": "replicate", "mode": "chat", + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "max_tokens": 163840, "supports_function_calling": true, + "supports_reasoning": true, "supports_system_messages": true }, "replicate/xai/grok-4": { @@ -30196,12 +26428,16 @@ "supports_function_calling": true, "supports_system_messages": true }, - "replicateopenai/gpt-oss-20b": { - "input_cost_per_token": 9e-08, - "output_cost_per_token": 3.6e-07, + "replicate/deepseek-ai/deepseek-r1": { + "input_cost_per_token": 3.75e-06, + "output_cost_per_token": 1e-05, + "output_cost_per_reasoning_token": 1e-05, "litellm_provider": "replicate", "mode": "chat", - "supports_function_calling": true, + "max_input_tokens": 65536, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_reasoning": true, "supports_system_messages": true }, "rerank-english-v2.0": { @@ -30259,116 +26495,26 @@ "mode": "rerank", "output_cost_per_token": 0.0 }, - "runwayml/eleven_multilingual_v2": { - "litellm_provider": "runwayml", - "mode": "audio_speech", - "input_cost_per_character": 3e-07, - "source": "https://docs.dev.runwayml.com/guides/pricing/", - "metadata": { - "comment": "Estimated cost based on standard TTS pricing. RunwayML uses ElevenLabs models." - } + "nvidia_nim/nvidia/nv-rerankqa-mistral-4b-v3": { + "input_cost_per_query": 0.0, + "input_cost_per_token": 0.0, + "litellm_provider": "nvidia_nim", + "mode": "rerank", + "output_cost_per_token": 0.0 }, - "runwayml/gen3a_turbo": { - "litellm_provider": "runwayml", - "mode": "video_generation", - "output_cost_per_video_per_second": 0.05, - "source": "https://docs.dev.runwayml.com/guides/pricing/", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "video" - ], - "supported_resolutions": [ - "1280x720", - "720x1280" - ], - "metadata": { - "comment": "5 credits per second @ $0.01 per credit = $0.05 per second" - } + "nvidia_nim/nvidia/llama-3_2-nv-rerankqa-1b-v2": { + "input_cost_per_query": 0.0, + "input_cost_per_token": 0.0, + "litellm_provider": "nvidia_nim", + "mode": "rerank", + "output_cost_per_token": 0.0 }, - "runwayml/gen4_aleph": { - "litellm_provider": "runwayml", - "mode": "video_generation", - "output_cost_per_video_per_second": 0.15, - "source": "https://docs.dev.runwayml.com/guides/pricing/", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "video" - ], - "supported_resolutions": [ - "1280x720", - "720x1280" - ], - "metadata": { - "comment": "15 credits per second @ $0.01 per credit = $0.15 per second" - } - }, - "runwayml/gen4_image": { - "litellm_provider": "runwayml", - "mode": "image_generation", - "input_cost_per_image": 0.05, - "output_cost_per_image": 0.05, - "source": "https://docs.dev.runwayml.com/guides/pricing/", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "image" - ], - "supported_resolutions": [ - "1280x720", - "1920x1080" - ], - "metadata": { - "comment": "5 credits per 720p image or 8 credits per 1080p image @ $0.01 per credit. Using 5 credits ($0.05) as base cost" - } - }, - "runwayml/gen4_image_turbo": { - "litellm_provider": "runwayml", - "mode": "image_generation", - "input_cost_per_image": 0.02, - "output_cost_per_image": 0.02, - "source": "https://docs.dev.runwayml.com/guides/pricing/", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "image" - ], - "supported_resolutions": [ - "1280x720", - "1920x1080" - ], - "metadata": { - "comment": "2 credits per image (any resolution) @ $0.01 per credit = $0.02 per image" - } - }, - "runwayml/gen4_turbo": { - "litellm_provider": "runwayml", - "mode": "video_generation", - "output_cost_per_video_per_second": 0.05, - "source": "https://docs.dev.runwayml.com/guides/pricing/", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "video" - ], - "supported_resolutions": [ - "1280x720", - "720x1280" - ], - "metadata": { - "comment": "5 credits per second @ $0.01 per credit = $0.05 per second" - } + "nvidia_nim/ranking/nvidia/llama-3.2-nv-rerankqa-1b-v2": { + "input_cost_per_query": 0.0, + "input_cost_per_token": 0.0, + "litellm_provider": "nvidia_nim", + "mode": "rerank", + "output_cost_per_token": 0.0 }, "sagemaker/meta-textgeneration-llama-2-13b": { "input_cost_per_token": 0.0, @@ -30457,19 +26603,6 @@ "supports_reasoning": true, "supports_tool_choice": true }, - "sambanova/DeepSeek-V3.1": { - "max_tokens": 32768, - "max_input_tokens": 32768, - "max_output_tokens": 32768, - "input_cost_per_token": 3e-06, - "output_cost_per_token": 4.5e-06, - "litellm_provider": "sambanova", - "mode": "chat", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_reasoning": true, - "source": "https://cloud.sambanova.ai/plans/pricing" - }, "sambanova/Llama-4-Maverick-17B-128E-Instruct": { "input_cost_per_token": 6.3e-07, "litellm_provider": "sambanova", @@ -30606,6 +26739,19 @@ "supports_reasoning": true, "supports_tool_choice": true }, + "sambanova/DeepSeek-V3.1": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 3e-06, + "output_cost_per_token": 4.5e-06, + "litellm_provider": "sambanova", + "mode": "chat", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_reasoning": true, + "source": "https://cloud.sambanova.ai/plans/pricing" + }, "sambanova/gpt-oss-120b": { "max_tokens": 131072, "max_input_tokens": 131072, @@ -30619,75 +26765,6 @@ "supports_reasoning": true, "source": "https://cloud.sambanova.ai/plans/pricing" }, - "sample_spec": { - "code_interpreter_cost_per_session": 0.0, - "computer_use_input_cost_per_1k_tokens": 0.0, - "computer_use_output_cost_per_1k_tokens": 0.0, - "deprecation_date": "date when the model becomes deprecated in the format YYYY-MM-DD", - "file_search_cost_per_1k_calls": 0.0, - "file_search_cost_per_gb_per_day": 0.0, - "input_cost_per_audio_token": 0.0, - "input_cost_per_token": 0.0, - "litellm_provider": "one of https://docs.litellm.ai/docs/providers", - "max_input_tokens": "max input tokens, if the provider specifies it. if not default to max_tokens", - "max_output_tokens": "max output tokens, if the provider specifies it. if not default to max_tokens", - "max_tokens": "LEGACY parameter. set to max_output_tokens if provider specifies it. IF not set to max_input_tokens, if provider specifies it.", - "mode": "one of: chat, embedding, completion, image_generation, audio_transcription, audio_speech, image_generation, moderation, rerank, search", - "output_cost_per_reasoning_token": 0.0, - "output_cost_per_token": 0.0, - "search_context_cost_per_query": { - "search_context_size_high": 0.0, - "search_context_size_low": 0.0, - "search_context_size_medium": 0.0 - }, - "supported_regions": [ - "global", - "us-west-2", - "eu-west-1", - "ap-southeast-1", - "ap-northeast-1" - ], - "supports_audio_input": true, - "supports_audio_output": true, - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_vision": true, - "supports_web_search": true, - "vector_store_cost_per_gb_per_day": 0.0 - }, - "sarvam/sarvam-m": { - "cache_creation_input_token_cost": 0, - "cache_creation_input_token_cost_above_1hr": 0, - "cache_read_input_token_cost": 0, - "input_cost_per_token": 0, - "litellm_provider": "sarvam", - "max_input_tokens": 8192, - "max_output_tokens": 32000, - "max_tokens": 32000, - "mode": "chat", - "output_cost_per_token": 0, - "supports_reasoning": true - }, - "searxng/search": { - "litellm_provider": "searxng", - "mode": "search", - "input_cost_per_query": 0.0, - "metadata": { - "notes": "SearXNG is an open-source metasearch engine. Free to use when self-hosted or using public instances." - } - }, - "serper/search": { - "input_cost_per_query": 0.001, - "litellm_provider": "serper", - "mode": "search", - "metadata": { - "notes": "Serper Google Search API. Pricing: $1.00/1k queries (Starter), $0.75/1k (Standard), $0.50/1k (Scale), $0.30/1k (Ultimate)." - } - }, "snowflake/claude-3-5-sonnet": { "litellm_provider": "snowflake", "max_input_tokens": 18000, @@ -30858,241 +26935,6 @@ "max_tokens": 8192, "mode": "chat" }, - "sora-2": { - "litellm_provider": "openai", - "mode": "video_generation", - "output_cost_per_video_per_second": 0.1, - "source": "https://platform.openai.com/docs/api-reference/videos", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "video" - ], - "supported_resolutions": [ - "720x1280", - "1280x720" - ] - }, - "sora-2-pro": { - "litellm_provider": "openai", - "mode": "video_generation", - "output_cost_per_video_per_second": 0.3, - "source": "https://platform.openai.com/docs/api-reference/videos", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "video" - ], - "supported_resolutions": [ - "720x1280", - "1280x720" - ] - }, - "sora-2-pro-high-res": { - "litellm_provider": "openai", - "mode": "video_generation", - "output_cost_per_video_per_second": 0.5, - "source": "https://platform.openai.com/docs/api-reference/videos", - "supported_modalities": [ - "text", - "image" - ], - "supported_output_modalities": [ - "video" - ], - "supported_resolutions": [ - "1024x1792", - "1792x1024" - ] - }, - "stability.sd3-5-large-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "max_tokens": 77, - "mode": "image_generation", - "output_cost_per_image": 0.08 - }, - "stability.sd3-large-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "max_tokens": 77, - "mode": "image_generation", - "output_cost_per_image": 0.08 - }, - "stability.stable-conservative-upscale-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.4 - }, - "stability.stable-creative-upscale-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.6 - }, - "stability.stable-fast-upscale-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.03 - }, - "stability.stable-image-control-sketch-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.07 - }, - "stability.stable-image-control-structure-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.07 - }, - "stability.stable-image-core-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "max_tokens": 77, - "mode": "image_generation", - "output_cost_per_image": 0.04 - }, - "stability.stable-image-core-v1:1": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "max_tokens": 77, - "mode": "image_generation", - "output_cost_per_image": 0.04 - }, - "stability.stable-image-erase-object-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.07 - }, - "stability.stable-image-inpaint-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.07 - }, - "stability.stable-image-remove-background-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.07 - }, - "stability.stable-image-search-recolor-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.07 - }, - "stability.stable-image-search-replace-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.07 - }, - "stability.stable-image-style-guide-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.07 - }, - "stability.stable-image-ultra-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "max_tokens": 77, - "mode": "image_generation", - "output_cost_per_image": 0.14 - }, - "stability.stable-image-ultra-v1:1": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "max_tokens": 77, - "mode": "image_generation", - "output_cost_per_image": 0.14 - }, - "stability.stable-outpaint-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.06 - }, - "stability.stable-style-transfer-v1:0": { - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "mode": "image_edit", - "output_cost_per_image": 0.08 - }, - "stability/conservative": { - "litellm_provider": "stability", - "mode": "image_edit", - "output_cost_per_image": 0.04, - "supported_endpoints": [ - "/v1/images/edits" - ] - }, - "stability/creative": { - "litellm_provider": "stability", - "mode": "image_edit", - "output_cost_per_image": 0.06, - "supported_endpoints": [ - "/v1/images/edits" - ] - }, - "stability/erase": { - "litellm_provider": "stability", - "mode": "image_edit", - "output_cost_per_image": 0.005, - "supported_endpoints": [ - "/v1/images/edits" - ] - }, - "stability/fast": { - "litellm_provider": "stability", - "mode": "image_edit", - "output_cost_per_image": 0.002, - "supported_endpoints": [ - "/v1/images/edits" - ] - }, - "stability/inpaint": { - "litellm_provider": "stability", - "mode": "image_edit", - "output_cost_per_image": 0.005, - "supported_endpoints": [ - "/v1/images/edits" - ] - }, - "stability/outpaint": { - "litellm_provider": "stability", - "mode": "image_edit", - "output_cost_per_image": 0.004, - "supported_endpoints": [ - "/v1/images/edits" - ] - }, - "stability/remove-background": { - "litellm_provider": "stability", - "mode": "image_edit", - "output_cost_per_image": 0.005, - "supported_endpoints": [ - "/v1/images/edits" - ] - }, - "stability/replace-background-and-relight": { - "litellm_provider": "stability", - "mode": "image_edit", - "output_cost_per_image": 0.008, - "supported_endpoints": [ - "/v1/images/edits" - ] - }, "stability/sd3": { "litellm_provider": "stability", "mode": "image_generation", @@ -31149,7 +26991,31 @@ "/v1/images/generations" ] }, - "stability/search-and-recolor": { + "stability/stable-image-ultra": { + "litellm_provider": "stability", + "mode": "image_generation", + "output_cost_per_image": 0.08, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "stability/inpaint": { + "litellm_provider": "stability", + "mode": "image_edit", + "output_cost_per_image": 0.005, + "supported_endpoints": [ + "/v1/images/edits" + ] + }, + "stability/outpaint": { + "litellm_provider": "stability", + "mode": "image_edit", + "output_cost_per_image": 0.004, + "supported_endpoints": [ + "/v1/images/edits" + ] + }, + "stability/erase": { "litellm_provider": "stability", "mode": "image_edit", "output_cost_per_image": 0.005, @@ -31165,7 +27031,7 @@ "/v1/images/edits" ] }, - "stability/sketch": { + "stability/search-and-recolor": { "litellm_provider": "stability", "mode": "image_edit", "output_cost_per_image": 0.005, @@ -31173,20 +27039,28 @@ "/v1/images/edits" ] }, - "stability/stable-image-core": { + "stability/remove-background": { "litellm_provider": "stability", - "mode": "image_generation", - "output_cost_per_image": 0.03, + "mode": "image_edit", + "output_cost_per_image": 0.005, "supported_endpoints": [ - "/v1/images/generations" + "/v1/images/edits" ] }, - "stability/stable-image-ultra": { + "stability/replace-background-and-relight": { "litellm_provider": "stability", - "mode": "image_generation", - "output_cost_per_image": 0.08, + "mode": "image_edit", + "output_cost_per_image": 0.008, "supported_endpoints": [ - "/v1/images/generations" + "/v1/images/edits" + ] + }, + "stability/sketch": { + "litellm_provider": "stability", + "mode": "image_edit", + "output_cost_per_image": 0.005, + "supported_endpoints": [ + "/v1/images/edits" ] }, "stability/structure": { @@ -31213,90 +27087,186 @@ "/v1/images/edits" ] }, + "stability/fast": { + "litellm_provider": "stability", + "mode": "image_edit", + "output_cost_per_image": 0.002, + "supported_endpoints": [ + "/v1/images/edits" + ] + }, + "stability/conservative": { + "litellm_provider": "stability", + "mode": "image_edit", + "output_cost_per_image": 0.04, + "supported_endpoints": [ + "/v1/images/edits" + ] + }, + "stability/creative": { + "litellm_provider": "stability", + "mode": "image_edit", + "output_cost_per_image": 0.06, + "supported_endpoints": [ + "/v1/images/edits" + ] + }, + "stability/stable-image-core": { + "litellm_provider": "stability", + "mode": "image_generation", + "output_cost_per_image": 0.03, + "supported_endpoints": [ + "/v1/images/generations" + ] + }, + "stability.sd3-5-large-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "max_tokens": 77, + "mode": "image_generation", + "output_cost_per_image": 0.08 + }, + "stability.sd3-large-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "max_tokens": 77, + "mode": "image_generation", + "output_cost_per_image": 0.08 + }, + "stability.stable-image-core-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "max_tokens": 77, + "mode": "image_generation", + "output_cost_per_image": 0.04 + }, + "stability.stable-conservative-upscale-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.4 + }, + "stability.stable-creative-upscale-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.6 + }, + "stability.stable-fast-upscale-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.03 + }, + "stability.stable-outpaint-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.06 + }, + "stability.stable-image-control-sketch-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.07 + }, + "stability.stable-image-control-structure-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.07 + }, + "stability.stable-image-erase-object-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.07 + }, + "stability.stable-image-inpaint-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.07 + }, + "stability.stable-image-remove-background-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.07 + }, + "stability.stable-image-search-recolor-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.07 + }, + "stability.stable-image-search-replace-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.07 + }, + "stability.stable-image-style-guide-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.07 + }, + "stability.stable-style-transfer-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "mode": "image_edit", + "output_cost_per_image": 0.08 + }, + "stability.stable-image-core-v1:1": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "max_tokens": 77, + "mode": "image_generation", + "output_cost_per_image": 0.04 + }, + "stability.stable-image-ultra-v1:0": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "max_tokens": 77, + "mode": "image_generation", + "output_cost_per_image": 0.14 + }, + "stability.stable-image-ultra-v1:1": { + "litellm_provider": "bedrock", + "max_input_tokens": 77, + "max_tokens": 77, + "mode": "image_generation", + "output_cost_per_image": 0.14 + }, "standard/1024-x-1024/dall-e-3": { "input_cost_per_pixel": 3.81469e-08, "litellm_provider": "openai", "mode": "image_generation", "output_cost_per_pixel": 0.0 }, - "standard/1024-x-1024/gpt-image-1.5": { - "input_cost_per_image": 0.009, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "standard/1024-x-1024/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.009, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "standard/1024-x-1536/gpt-image-1.5": { - "input_cost_per_image": 0.013, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "standard/1024-x-1536/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.013, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, "standard/1024-x-1792/dall-e-3": { "input_cost_per_pixel": 4.359e-08, "litellm_provider": "openai", "mode": "image_generation", "output_cost_per_pixel": 0.0 }, - "standard/1536-x-1024/gpt-image-1.5": { - "input_cost_per_image": 0.013, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, - "standard/1536-x-1024/gpt-image-1.5-2025-12-16": { - "input_cost_per_image": 0.013, - "litellm_provider": "openai", - "mode": "image_generation", - "supported_endpoints": [ - "/v1/images/generations", - "/v1/images/edits" - ], - "supports_vision": true, - "supports_pdf_input": true - }, "standard/1792-x-1024/dall-e-3": { "input_cost_per_pixel": 4.359e-08, "litellm_provider": "openai", "mode": "image_generation", "output_cost_per_pixel": 0.0 }, + "linkup/search": { + "input_cost_per_query": 0.00587, + "litellm_provider": "linkup", + "mode": "search" + }, + "linkup/search-deep": { + "input_cost_per_query": 0.05867, + "litellm_provider": "linkup", + "mode": "search" + }, "tavily/search": { "input_cost_per_query": 0.008, "litellm_provider": "tavily", @@ -31514,11 +27484,13 @@ "mode": "embedding", "output_cost_per_token": 0.0 }, - "together-ai-up-to-4b": { - "input_cost_per_token": 1e-07, + "together_ai/baai/bge-base-en-v1.5": { + "input_cost_per_token": 8e-09, "litellm_provider": "together_ai", - "mode": "chat", - "output_cost_per_token": 1e-07 + "max_input_tokens": 512, + "mode": "embedding", + "output_cost_per_token": 0.0, + "output_vector_size": 768 }, "together_ai/BAAI/bge-base-en-v1.5": { "input_cost_per_token": 8e-09, @@ -31528,6 +27500,12 @@ "output_cost_per_token": 0.0, "output_vector_size": 768 }, + "together-ai-up-to-4b": { + "input_cost_per_token": 1e-07, + "litellm_provider": "together_ai", + "mode": "chat", + "output_cost_per_token": 1e-07 + }, "together_ai/Qwen/Qwen2.5-72B-Instruct-Turbo": { "litellm_provider": "together_ai", "mode": "chat", @@ -31591,50 +27569,6 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "together_ai/Qwen/Qwen3-Next-80B-A3B-Instruct": { - "input_cost_per_token": 1.5e-07, - "litellm_provider": "together_ai", - "max_input_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 1.5e-06, - "source": "https://www.together.ai/models/qwen3-next-80b-a3b-instruct", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "together_ai/Qwen/Qwen3-Next-80B-A3B-Thinking": { - "input_cost_per_token": 1.5e-07, - "litellm_provider": "together_ai", - "max_input_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 1.5e-06, - "source": "https://www.together.ai/models/qwen3-next-80b-a3b-thinking", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "together_ai/Qwen/Qwen3.5-397B-A17B": { - "input_cost_per_token": 6e-07, - "litellm_provider": "together_ai", - "max_input_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 3.6e-06, - "source": "https://www.together.ai/models/Qwen/Qwen3.5-397B-A17B", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_response_schema": true, - "supports_tool_choice": true - }, - "together_ai/baai/bge-base-en-v1.5": { - "input_cost_per_token": 8e-09, - "litellm_provider": "together_ai", - "max_input_tokens": 512, - "mode": "embedding", - "output_cost_per_token": 0.0, - "output_vector_size": 768 - }, "together_ai/deepseek-ai/DeepSeek-R1": { "input_cost_per_token": 3e-06, "litellm_provider": "together_ai", @@ -31799,31 +27733,6 @@ "supports_response_schema": true, "supports_tool_choice": true }, - "together_ai/moonshotai/Kimi-K2-Instruct-0905": { - "input_cost_per_token": 1e-06, - "litellm_provider": "together_ai", - "max_input_tokens": 262144, - "mode": "chat", - "output_cost_per_token": 3e-06, - "source": "https://www.together.ai/models/kimi-k2-0905", - "supports_function_calling": true, - "supports_parallel_function_calling": true, - "supports_tool_choice": true - }, - "together_ai/moonshotai/Kimi-K2.5": { - "input_cost_per_token": 5e-07, - "litellm_provider": "together_ai", - "max_input_tokens": 256000, - "max_output_tokens": 256000, - "max_tokens": 256000, - "mode": "chat", - "output_cost_per_token": 2.8e-06, - "source": "https://www.together.ai/models/kimi-k2-5", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_reasoning": true - }, "together_ai/openai/gpt-oss-120b": { "input_cost_per_token": 1.5e-07, "litellm_provider": "together_ai", @@ -31895,15 +27804,68 @@ "supports_reasoning": true, "supports_tool_choice": true }, - "tts-1": { - "input_cost_per_character": 1.5e-05, - "litellm_provider": "openai", - "mode": "audio_speech", - "supported_endpoints": [ - "/v1/audio/speech" - ] + "together_ai/moonshotai/Kimi-K2.5": { + "input_cost_per_token": 5e-07, + "litellm_provider": "together_ai", + "max_input_tokens": 256000, + "max_output_tokens": 256000, + "max_tokens": 256000, + "mode": "chat", + "output_cost_per_token": 2.8e-06, + "source": "https://www.together.ai/models/kimi-k2-5", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_reasoning": true }, - "tts-1-1106": { + "together_ai/moonshotai/Kimi-K2-Instruct-0905": { + "input_cost_per_token": 1e-06, + "litellm_provider": "together_ai", + "max_input_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 3e-06, + "source": "https://www.together.ai/models/kimi-k2-0905", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true + }, + "together_ai/Qwen/Qwen3-Next-80B-A3B-Instruct": { + "input_cost_per_token": 1.5e-07, + "litellm_provider": "together_ai", + "max_input_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 1.5e-06, + "source": "https://www.together.ai/models/qwen3-next-80b-a3b-instruct", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "together_ai/Qwen/Qwen3-Next-80B-A3B-Thinking": { + "input_cost_per_token": 1.5e-07, + "litellm_provider": "together_ai", + "max_input_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 1.5e-06, + "source": "https://www.together.ai/models/qwen3-next-80b-a3b-thinking", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "together_ai/Qwen/Qwen3.5-397B-A17B": { + "input_cost_per_token": 6e-07, + "litellm_provider": "together_ai", + "max_input_tokens": 262144, + "mode": "chat", + "output_cost_per_token": 3.6e-06, + "source": "https://www.together.ai/models/Qwen/Qwen3.5-397B-A17B", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "tts-1": { "input_cost_per_character": 1.5e-05, "litellm_provider": "openai", "mode": "audio_speech", @@ -31919,67 +27881,41 @@ "/v1/audio/speech" ] }, - "tts-1-hd-1106": { - "input_cost_per_character": 3e-05, - "litellm_provider": "openai", + "aws_polly/standard": { + "input_cost_per_character": 4e-06, + "litellm_provider": "aws_polly", "mode": "audio_speech", "supported_endpoints": [ "/v1/audio/speech" - ] + ], + "source": "https://aws.amazon.com/polly/pricing/" }, - "twelvelabs.marengo-embed-2-7-v1:0": { - "input_cost_per_token": 7e-05, - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "max_tokens": 77, - "mode": "embedding", - "output_cost_per_token": 0.0, - "output_vector_size": 1024, - "supports_embedding_image_input": true, - "supports_image_input": true + "aws_polly/neural": { + "input_cost_per_character": 1.6e-05, + "litellm_provider": "aws_polly", + "mode": "audio_speech", + "supported_endpoints": [ + "/v1/audio/speech" + ], + "source": "https://aws.amazon.com/polly/pricing/" }, - "twelvelabs.pegasus-1-2-v1:0": { - "input_cost_per_video_per_second": 0.00049, - "output_cost_per_token": 7.5e-06, - "litellm_provider": "bedrock", - "mode": "chat", - "supports_video_input": true + "aws_polly/long-form": { + "input_cost_per_character": 0.0001, + "litellm_provider": "aws_polly", + "mode": "audio_speech", + "supported_endpoints": [ + "/v1/audio/speech" + ], + "source": "https://aws.amazon.com/polly/pricing/" }, - "us.amazon.nova-2-lite-v1:0": { - "cache_read_input_token_cost": 8.25e-08, - "input_cost_per_token": 3.3e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 2.75e-06, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_video_input": true, - "supports_vision": true - }, - "us.amazon.nova-2-pro-preview-20251202-v1:0": { - "cache_read_input_token_cost": 5.46875e-07, - "input_cost_per_token": 2.1875e-06, - "input_cost_per_image_token": 2.1875e-06, - "input_cost_per_audio_token": 2.1875e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.75e-05, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_video_input": true, - "supports_vision": true + "aws_polly/generative": { + "input_cost_per_character": 3e-05, + "litellm_provider": "aws_polly", + "mode": "audio_speech", + "supported_endpoints": [ + "/v1/audio/speech" + ], + "source": "https://aws.amazon.com/polly/pricing/" }, "us.amazon.nova-lite-v1:0": { "input_cost_per_token": 6e-08, @@ -32052,6 +27988,28 @@ "supports_response_schema": true, "supports_tool_choice": true }, + "us.anthropic.claude-haiku-4-5-20251001-v1:0": { + "cache_creation_input_token_cost": 1.375e-06, + "cache_read_input_token_cost": 1.1e-07, + "input_cost_per_token": 1.1e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 5.5e-06, + "source": "https://aws.amazon.com/about-aws/whats-new/2025/10/claude-4-5-haiku-anthropic-amazon-bedrock", + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, "us.anthropic.claude-3-5-sonnet-20240620-v1:0": { "input_cost_per_token": 3e-06, "litellm_provider": "bedrock", @@ -32154,28 +28112,6 @@ "cache_read_input_token_cost": 3e-07, "cache_creation_input_token_cost": 3.75e-06 }, - "us.anthropic.claude-haiku-4-5-20251001-v1:0": { - "cache_creation_input_token_cost": 1.375e-06, - "cache_read_input_token_cost": 1.1e-07, - "input_cost_per_token": 1.1e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 5.5e-06, - "source": "https://aws.amazon.com/about-aws/whats-new/2025/10/claude-4-5-haiku-anthropic-amazon-bedrock", - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, "us.anthropic.claude-opus-4-1-20250805-v1:0": { "cache_creation_input_token_cost": 1.875e-05, "cache_read_input_token_cost": 1.5e-06, @@ -32202,6 +28138,57 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, + "us.anthropic.claude-sonnet-4-5-20250929-v1:0": { + "cache_creation_input_token_cost": 4.125e-06, + "cache_read_input_token_cost": 3.3e-07, + "input_cost_per_token": 3.3e-06, + "input_cost_per_token_above_200k_tokens": 6.6e-06, + "output_cost_per_token_above_200k_tokens": 2.475e-05, + "cache_creation_input_token_cost_above_200k_tokens": 8.25e-06, + "cache_read_input_token_cost_above_200k_tokens": 6.6e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.65e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, + "au.anthropic.claude-haiku-4-5-20251001-v1:0": { + "cache_creation_input_token_cost": 1.375e-06, + "cache_read_input_token_cost": 1.1e-07, + "input_cost_per_token": 1.1e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 5.5e-06, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346 + }, "us.anthropic.claude-opus-4-20250514-v1:0": { "cache_creation_input_token_cost": 1.875e-05, "cache_read_input_token_cost": 1.5e-06, @@ -32254,26 +28241,22 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, - "us.anthropic.claude-opus-4-6-v1": { - "cache_creation_input_token_cost": 6.875e-06, - "cache_creation_input_token_cost_above_200k_tokens": 1.375e-05, - "cache_read_input_token_cost": 5.5e-07, - "cache_read_input_token_cost_above_200k_tokens": 1.1e-06, - "input_cost_per_token": 5.5e-06, - "input_cost_per_token_above_200k_tokens": 1.1e-05, + "global.anthropic.claude-opus-4-5-20251101-v1:0": { + "cache_creation_input_token_cost": 6.25e-06, + "cache_read_input_token_cost": 5e-07, + "input_cost_per_token": 5e-06, "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 128000, - "max_tokens": 128000, + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, "mode": "chat", - "output_cost_per_token": 2.75e-05, - "output_cost_per_token_above_200k_tokens": 4.125e-05, + "output_cost_per_token": 2.5e-05, "search_context_cost_per_query": { "search_context_size_high": 0.01, "search_context_size_low": 0.01, "search_context_size_medium": 0.01 }, - "supports_assistant_prefill": false, + "supports_assistant_prefill": true, "supports_computer_use": true, "supports_function_calling": true, "supports_pdf_input": true, @@ -32282,7 +28265,33 @@ "supports_response_schema": true, "supports_tool_choice": true, "supports_vision": true, - "tool_use_system_prompt_tokens": 346 + "tool_use_system_prompt_tokens": 159 + }, + "eu.anthropic.claude-opus-4-5-20251101-v1:0": { + "cache_creation_input_token_cost": 6.25e-06, + "cache_read_input_token_cost": 5e-07, + "input_cost_per_token": 5e-06, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 2.5e-05, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + }, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 159 }, "us.anthropic.claude-sonnet-4-20250514-v1:0": { "cache_creation_input_token_cost": 3.75e-06, @@ -32314,66 +28323,6 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, - "us.anthropic.claude-sonnet-4-5-20250929-v1:0": { - "cache_creation_input_token_cost": 4.125e-06, - "cache_read_input_token_cost": 3.3e-07, - "input_cost_per_token": 3.3e-06, - "input_cost_per_token_above_200k_tokens": 6.6e-06, - "output_cost_per_token_above_200k_tokens": 2.475e-05, - "cache_creation_input_token_cost_above_200k_tokens": 8.25e-06, - "cache_read_input_token_cost_above_200k_tokens": 6.6e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.65e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, - "us.anthropic.claude-sonnet-4-6": { - "cache_creation_input_token_cost": 4.125e-06, - "cache_creation_input_token_cost_above_200k_tokens": 8.25e-06, - "cache_read_input_token_cost": 3.3e-07, - "cache_read_input_token_cost_above_200k_tokens": 6.6e-07, - "input_cost_per_token": 3.3e-06, - "input_cost_per_token_above_200k_tokens": 6.6e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.65e-05, - "output_cost_per_token_above_200k_tokens": 2.475e-05, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - }, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346 - }, "us.deepseek.r1-v1:0": { "input_cost_per_token": 1.35e-06, "litellm_provider": "bedrock_converse", @@ -32398,6 +28347,18 @@ "supports_reasoning": true, "supports_tool_choice": true }, + "eu.deepseek.v3.2": { + "input_cost_per_token": 7.4e-07, + "litellm_provider": "bedrock_converse", + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "max_tokens": 163840, + "mode": "chat", + "output_cost_per_token": 2.22e-06, + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, "us.meta.llama3-1-405b-instruct-v1:0": { "input_cost_per_token": 5.32e-06, "litellm_provider": "bedrock", @@ -32541,49 +28502,6 @@ "supports_function_calling": true, "supports_tool_choice": false }, - "us.twelvelabs.marengo-embed-2-7-v1:0": { - "input_cost_per_token": 7e-05, - "input_cost_per_video_per_second": 0.0007, - "input_cost_per_audio_per_second": 0.00014, - "input_cost_per_image": 0.0001, - "litellm_provider": "bedrock", - "max_input_tokens": 77, - "max_tokens": 77, - "mode": "embedding", - "output_cost_per_token": 0.0, - "output_vector_size": 1024, - "supports_embedding_image_input": true, - "supports_image_input": true - }, - "us.twelvelabs.pegasus-1-2-v1:0": { - "input_cost_per_video_per_second": 0.00049, - "output_cost_per_token": 7.5e-06, - "litellm_provider": "bedrock", - "mode": "chat", - "supports_video_input": true - }, - "us.writer.palmyra-x4-v1:0": { - "input_cost_per_token": 2.5e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 128000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1e-05, - "supports_function_calling": true, - "supports_pdf_input": true - }, - "us.writer.palmyra-x5-v1:0": { - "input_cost_per_token": 6e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 6e-06, - "supports_function_calling": true, - "supports_pdf_input": true - }, "v0/v0-1.0-md": { "input_cost_per_token": 3e-06, "litellm_provider": "v0", @@ -32719,61 +28637,6 @@ "mode": "chat", "output_cost_per_token": 0.0 }, - "vercel_ai_gateway/anthropic/claude-3-5-sonnet": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 3e-06, - "litellm_provider": "vercel_ai_gateway", - "max_input_tokens": 200000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true - }, - "vercel_ai_gateway/anthropic/claude-3-5-sonnet-20241022": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 3e-06, - "litellm_provider": "vercel_ai_gateway", - "max_input_tokens": 200000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true - }, - "vercel_ai_gateway/anthropic/claude-3-7-sonnet": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 3e-06, - "litellm_provider": "vercel_ai_gateway", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true - }, "vercel_ai_gateway/anthropic/claude-3-haiku": { "cache_creation_input_token_cost": 3e-07, "cache_read_input_token_cost": 3e-08, @@ -32877,6 +28740,61 @@ "supports_function_calling": true, "supports_tool_choice": true }, + "vercel_ai_gateway/anthropic/claude-3-5-sonnet": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 3e-06, + "litellm_provider": "vercel_ai_gateway", + "max_input_tokens": 200000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "vercel_ai_gateway/anthropic/claude-3-5-sonnet-20241022": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 3e-06, + "litellm_provider": "vercel_ai_gateway", + "max_input_tokens": 200000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "vercel_ai_gateway/anthropic/claude-3-7-sonnet": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 3e-06, + "litellm_provider": "vercel_ai_gateway", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true + }, "vercel_ai_gateway/anthropic/claude-haiku-4.5": { "cache_creation_input_token_cost": 1.25e-06, "cache_read_input_token_cost": 1e-07, @@ -33888,6 +29806,27 @@ "supports_pdf_input": true, "supports_tool_choice": true }, + "vertex_ai/claude-haiku-4-5@20251001": { + "cache_creation_input_token_cost": 1.25e-06, + "cache_read_input_token_cost": 1e-07, + "input_cost_per_token": 1e-06, + "litellm_provider": "vertex_ai-anthropic_models", + "max_input_tokens": 200000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 5e-06, + "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude/haiku-4-5", + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_native_streaming": true, + "supports_vision": true + }, "vertex_ai/claude-3-5-sonnet": { "input_cost_per_token": 3e-06, "litellm_provider": "vertex_ai-anthropic_models", @@ -34025,27 +29964,6 @@ "supports_tool_choice": true, "supports_vision": true }, - "vertex_ai/claude-haiku-4-5@20251001": { - "cache_creation_input_token_cost": 1.25e-06, - "cache_read_input_token_cost": 1e-07, - "input_cost_per_token": 1e-06, - "litellm_provider": "vertex_ai-anthropic_models", - "max_input_tokens": 200000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 5e-06, - "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude/haiku-4-5", - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_native_streaming": true, - "supports_vision": true - }, "vertex_ai/claude-opus-4": { "cache_creation_input_token_cost": 1.875e-05, "cache_read_input_token_cost": 1.5e-06, @@ -34219,6 +30137,89 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 346 }, + "vertex_ai/claude-sonnet-4-5": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, + "input_cost_per_token_batches": 1.5e-06, + "litellm_provider": "vertex_ai-anthropic_models", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "output_cost_per_token_batches": 7.5e-06, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true + }, + "vertex_ai/claude-sonnet-4-6": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost": 3e-07, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "litellm_provider": "vertex_ai-anthropic_models", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + } + }, + "vertex_ai/claude-sonnet-4-5@20250929": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, + "input_cost_per_token_batches": 1.5e-06, + "litellm_provider": "vertex_ai-anthropic_models", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "output_cost_per_token_batches": 7.5e-06, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_native_streaming": true + }, "vertex_ai/claude-opus-4@20250514": { "cache_creation_input_token_cost": 1.875e-05, "cache_read_input_token_cost": 1.5e-06, @@ -34275,119 +30276,6 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, - "vertex_ai/claude-sonnet-4-5": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "input_cost_per_token_batches": 1.5e-06, - "litellm_provider": "vertex_ai-anthropic_models", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "output_cost_per_token_batches": 7.5e-06, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true - }, - "vertex_ai/claude-sonnet-4-5@20250929": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "input_cost_per_token_batches": 1.5e-06, - "litellm_provider": "vertex_ai-anthropic_models", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "output_cost_per_token_batches": 7.5e-06, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_native_streaming": true - }, - "vertex_ai/claude-sonnet-4-6": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost": 3e-07, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "litellm_provider": "vertex_ai-anthropic_models", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - } - }, - "vertex_ai/claude-sonnet-4-6@default": { - "cache_creation_input_token_cost": 3.75e-06, - "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, - "cache_read_input_token_cost": 3e-07, - "cache_read_input_token_cost_above_200k_tokens": 6e-07, - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_200k_tokens": 6e-06, - "litellm_provider": "vertex_ai-anthropic_models", - "max_input_tokens": 200000, - "max_output_tokens": 64000, - "max_tokens": 64000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "output_cost_per_token_above_200k_tokens": 2.25e-05, - "supports_assistant_prefill": true, - "supports_computer_use": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "tool_use_system_prompt_tokens": 346, - "search_context_cost_per_query": { - "search_context_size_high": 0.01, - "search_context_size_low": 0.01, - "search_context_size_medium": 0.01 - } - }, "vertex_ai/claude-sonnet-4@20250514": { "cache_creation_input_token_cost": 3.75e-06, "cache_read_input_token_cost": 3e-07, @@ -34418,6 +30306,17 @@ "supports_vision": true, "tool_use_system_prompt_tokens": 159 }, + "vertex_ai/mistralai/codestral-2@001": { + "input_cost_per_token": 3e-07, + "litellm_provider": "vertex_ai-mistral_models", + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 9e-07, + "supports_function_calling": true, + "supports_tool_choice": true + }, "vertex_ai/codestral-2": { "input_cost_per_token": 3e-07, "litellm_provider": "vertex_ai-mistral_models", @@ -34429,6 +30328,28 @@ "supports_function_calling": true, "supports_tool_choice": true }, + "vertex_ai/codestral-2@001": { + "input_cost_per_token": 3e-07, + "litellm_provider": "vertex_ai-mistral_models", + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 9e-07, + "supports_function_calling": true, + "supports_tool_choice": true + }, + "vertex_ai/mistralai/codestral-2": { + "input_cost_per_token": 3e-07, + "litellm_provider": "vertex_ai-mistral_models", + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 9e-07, + "supports_function_calling": true, + "supports_tool_choice": true + }, "vertex_ai/codestral-2501": { "input_cost_per_token": 2e-07, "litellm_provider": "vertex_ai-mistral_models", @@ -34440,17 +30361,6 @@ "supports_function_calling": true, "supports_tool_choice": true }, - "vertex_ai/codestral-2@001": { - "input_cost_per_token": 3e-07, - "litellm_provider": "vertex_ai-mistral_models", - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 9e-07, - "supports_function_calling": true, - "supports_tool_choice": true - }, "vertex_ai/codestral@2405": { "input_cost_per_token": 2e-07, "litellm_provider": "vertex_ai-mistral_models", @@ -34473,44 +30383,6 @@ "supports_function_calling": true, "supports_tool_choice": true }, - "vertex_ai/deep-research-pro-preview-12-2025": { - "input_cost_per_image": 0.0011, - "input_cost_per_token": 2e-06, - "input_cost_per_token_batches": 1e-06, - "litellm_provider": "vertex_ai-language-models", - "max_input_tokens": 65536, - "max_output_tokens": 32768, - "max_tokens": 32768, - "mode": "image_generation", - "output_cost_per_image": 0.134, - "output_cost_per_image_token": 0.00012, - "output_cost_per_token": 1.2e-05, - "output_cost_per_token_batches": 6e-06, - "source": "https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/3-pro-image" - }, - "vertex_ai/deepseek-ai/deepseek-ocr-maas": { - "litellm_provider": "vertex_ai", - "mode": "ocr", - "input_cost_per_token": 3e-07, - "output_cost_per_token": 1.2e-06, - "ocr_cost_per_page": 0.0003, - "source": "https://cloud.google.com/vertex-ai/pricing" - }, - "vertex_ai/deepseek-ai/deepseek-r1-0528-maas": { - "input_cost_per_token": 1.35e-06, - "litellm_provider": "vertex_ai-deepseek_models", - "max_input_tokens": 65336, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 5.4e-06, - "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#partner-models", - "supports_assistant_prefill": true, - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, "vertex_ai/deepseek-ai/deepseek-v3.1-maas": { "input_cost_per_token": 1.35e-06, "litellm_provider": "vertex_ai-deepseek_models", @@ -34549,6 +30421,21 @@ "supports_reasoning": true, "supports_tool_choice": true }, + "vertex_ai/deepseek-ai/deepseek-r1-0528-maas": { + "input_cost_per_token": 1.35e-06, + "litellm_provider": "vertex_ai-deepseek_models", + "max_input_tokens": 65336, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 5.4e-06, + "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#partner-models", + "supports_assistant_prefill": true, + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, "vertex_ai/gemini-2.5-flash-image": { "cache_read_input_token_cost": 3e-08, "input_cost_per_audio_token": 1e-06, @@ -34598,55 +30485,6 @@ "supports_web_search": false, "tpm": 8000000 }, - "vertex_ai/gemini-3-flash-preview": { - "cache_read_input_token_cost": 5e-08, - "input_cost_per_token": 5e-07, - "input_cost_per_audio_token": 1e-06, - "litellm_provider": "vertex_ai", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_token": 3e-06, - "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_input": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_video_input": true, - "supports_vision": true, - "supports_web_search": true, - "supports_native_streaming": true, - "input_cost_per_token_priority": 9e-07, - "input_cost_per_audio_token_priority": 1.8e-06, - "output_cost_per_token_priority": 5.4e-06, - "cache_read_input_token_cost_priority": 9e-08, - "supports_service_tier": true - }, "vertex_ai/gemini-3-pro-image-preview": { "input_cost_per_image": 0.0011, "input_cost_per_token": 2e-06, @@ -34662,62 +30500,6 @@ "output_cost_per_token_batches": 6e-06, "source": "https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/3-pro-image" }, - "vertex_ai/gemini-3-pro-preview": { - "cache_read_input_token_cost": 2e-07, - "cache_read_input_token_cost_above_200k_tokens": 4e-07, - "cache_creation_input_token_cost_above_200k_tokens": 2.5e-07, - "input_cost_per_token": 2e-06, - "input_cost_per_token_above_200k_tokens": 4e-06, - "input_cost_per_token_batches": 1e-06, - "litellm_provider": "vertex_ai", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65535, - "max_pdf_size_mb": 30, - "max_tokens": 65535, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_token": 1.2e-05, - "output_cost_per_token_above_200k_tokens": 1.8e-05, - "output_cost_per_token_batches": 6e-06, - "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_input": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_video_input": true, - "supports_vision": true, - "supports_web_search": true, - "supports_native_streaming": true, - "input_cost_per_token_priority": 3.6e-06, - "input_cost_per_token_above_200k_tokens_priority": 7.2e-06, - "output_cost_per_token_priority": 2.16e-05, - "output_cost_per_token_above_200k_tokens_priority": 3.24e-05, - "cache_read_input_token_cost_priority": 3.6e-07, - "cache_read_input_token_cost_above_200k_tokens_priority": 7.2e-07, - "supports_service_tier": true - }, "vertex_ai/gemini-3.1-flash-image-preview": { "input_cost_per_image": 0.00056, "input_cost_per_token": 5e-07, @@ -34782,121 +30564,20 @@ "supports_web_search": true, "supports_native_streaming": true }, - "vertex_ai/gemini-3.1-pro-preview": { - "cache_read_input_token_cost": 2e-07, - "cache_read_input_token_cost_above_200k_tokens": 4e-07, - "cache_creation_input_token_cost_above_200k_tokens": 2.5e-07, + "vertex_ai/deep-research-pro-preview-12-2025": { + "input_cost_per_image": 0.0011, "input_cost_per_token": 2e-06, - "input_cost_per_token_above_200k_tokens": 4e-06, "input_cost_per_token_batches": 1e-06, - "litellm_provider": "vertex_ai", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65536, - "max_pdf_size_mb": 30, - "max_tokens": 65536, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", + "litellm_provider": "vertex_ai-language-models", + "max_input_tokens": 65536, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "image_generation", + "output_cost_per_image": 0.134, + "output_cost_per_image_token": 0.00012, "output_cost_per_token": 1.2e-05, - "output_cost_per_token_above_200k_tokens": 1.8e-05, "output_cost_per_token_batches": 6e-06, - "output_cost_per_image": 0.00012, - "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#gemini-models", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_input": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_video_input": true, - "supports_vision": true, - "supports_web_search": true, - "supports_url_context": true, - "supports_native_streaming": true, - "input_cost_per_token_priority": 3.6e-06, - "input_cost_per_token_above_200k_tokens_priority": 7.2e-06, - "output_cost_per_token_priority": 2.16e-05, - "output_cost_per_token_above_200k_tokens_priority": 3.24e-05, - "cache_read_input_token_cost_priority": 3.6e-07, - "cache_read_input_token_cost_above_200k_tokens_priority": 7.2e-07, - "supports_service_tier": true - }, - "vertex_ai/gemini-3.1-pro-preview-customtools": { - "cache_read_input_token_cost": 2e-07, - "cache_read_input_token_cost_above_200k_tokens": 4e-07, - "cache_creation_input_token_cost_above_200k_tokens": 2.5e-07, - "input_cost_per_token": 2e-06, - "input_cost_per_token_above_200k_tokens": 4e-06, - "input_cost_per_token_batches": 1e-06, - "litellm_provider": "vertex_ai", - "max_audio_length_hours": 8.4, - "max_audio_per_prompt": 1, - "max_images_per_prompt": 3000, - "max_input_tokens": 1048576, - "max_output_tokens": 65536, - "max_pdf_size_mb": 30, - "max_tokens": 65536, - "max_video_length": 1, - "max_videos_per_prompt": 10, - "mode": "chat", - "output_cost_per_token": 1.2e-05, - "output_cost_per_token_above_200k_tokens": 1.8e-05, - "output_cost_per_token_batches": 6e-06, - "output_cost_per_image": 0.00012, - "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#gemini-models", - "supported_endpoints": [ - "/v1/chat/completions", - "/v1/completions", - "/v1/batch" - ], - "supported_modalities": [ - "text", - "image", - "audio", - "video" - ], - "supported_output_modalities": [ - "text" - ], - "supports_audio_input": true, - "supports_function_calling": true, - "supports_pdf_input": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_video_input": true, - "supports_vision": true, - "supports_web_search": true, - "supports_url_context": true, - "supports_native_streaming": true, - "input_cost_per_token_priority": 3.6e-06, - "input_cost_per_token_above_200k_tokens_priority": 7.2e-06, - "output_cost_per_token_priority": 2.16e-05, - "output_cost_per_token_above_200k_tokens_priority": 3.24e-05, - "cache_read_input_token_cost_priority": 3.6e-07, - "cache_read_input_token_cost_above_200k_tokens_priority": 7.2e-07, - "supports_service_tier": true + "source": "https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/3-pro-image" }, "vertex_ai/imagegeneration@006": { "litellm_provider": "vertex_ai-image-models", @@ -34904,12 +30585,6 @@ "output_cost_per_image": 0.02, "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing" }, - "vertex_ai/imagen-3.0-capability-001": { - "litellm_provider": "vertex_ai-image-models", - "mode": "image_generation", - "output_cost_per_image": 0.04, - "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/image/edit-insert-objects" - }, "vertex_ai/imagen-3.0-fast-generate-001": { "litellm_provider": "vertex_ai-image-models", "mode": "image_generation", @@ -34929,6 +30604,12 @@ "output_cost_per_image": 0.04, "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing" }, + "vertex_ai/imagen-3.0-capability-001": { + "litellm_provider": "vertex_ai-image-models", + "mode": "image_generation", + "output_cost_per_image": 0.04, + "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/image/edit-insert-objects" + }, "vertex_ai/imagen-4.0-fast-generate-001": { "litellm_provider": "vertex_ai-image-models", "mode": "image_generation", @@ -35180,6 +30861,91 @@ "supports_function_calling": true, "supports_tool_choice": true }, + "vertex_ai/moonshotai/kimi-k2-thinking-maas": { + "input_cost_per_token": 6e-07, + "litellm_provider": "vertex_ai-moonshot_models", + "max_input_tokens": 256000, + "max_output_tokens": 256000, + "max_tokens": 256000, + "mode": "chat", + "output_cost_per_token": 2.5e-06, + "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#partner-models", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_web_search": true + }, + "vertex_ai/zai-org/glm-4.7-maas": { + "input_cost_per_token": 6e-07, + "litellm_provider": "vertex_ai-zai_models", + "max_input_tokens": 200000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 2.2e-06, + "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#partner-models", + "supports_function_calling": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, + "vertex_ai/zai-org/glm-5-maas": { + "cache_read_input_token_cost": 1e-07, + "input_cost_per_token": 1e-06, + "litellm_provider": "vertex_ai-zai_models", + "max_input_tokens": 200000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 3.2e-06, + "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#glm-models", + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true + }, + "vertex_ai/mistral-medium-3": { + "input_cost_per_token": 4e-07, + "litellm_provider": "vertex_ai-mistral_models", + "max_input_tokens": 128000, + "max_output_tokens": 8191, + "max_tokens": 8191, + "mode": "chat", + "output_cost_per_token": 2e-06, + "supports_function_calling": true, + "supports_tool_choice": true + }, + "vertex_ai/mistral-medium-3@001": { + "input_cost_per_token": 4e-07, + "litellm_provider": "vertex_ai-mistral_models", + "max_input_tokens": 128000, + "max_output_tokens": 8191, + "max_tokens": 8191, + "mode": "chat", + "output_cost_per_token": 2e-06, + "supports_function_calling": true, + "supports_tool_choice": true + }, + "vertex_ai/mistralai/mistral-medium-3": { + "input_cost_per_token": 4e-07, + "litellm_provider": "vertex_ai-mistral_models", + "max_input_tokens": 128000, + "max_output_tokens": 8191, + "max_tokens": 8191, + "mode": "chat", + "output_cost_per_token": 2e-06, + "supports_function_calling": true, + "supports_tool_choice": true + }, + "vertex_ai/mistralai/mistral-medium-3@001": { + "input_cost_per_token": 4e-07, + "litellm_provider": "vertex_ai-mistral_models", + "max_input_tokens": 128000, + "max_output_tokens": 8191, + "max_tokens": 8191, + "mode": "chat", + "output_cost_per_token": 2e-06, + "supports_function_calling": true, + "supports_tool_choice": true + }, "vertex_ai/mistral-large-2411": { "input_cost_per_token": 2e-06, "litellm_provider": "vertex_ai-mistral_models", @@ -35224,28 +30990,6 @@ "supports_function_calling": true, "supports_tool_choice": true }, - "vertex_ai/mistral-medium-3": { - "input_cost_per_token": 4e-07, - "litellm_provider": "vertex_ai-mistral_models", - "max_input_tokens": 128000, - "max_output_tokens": 8191, - "max_tokens": 8191, - "mode": "chat", - "output_cost_per_token": 2e-06, - "supports_function_calling": true, - "supports_tool_choice": true - }, - "vertex_ai/mistral-medium-3@001": { - "input_cost_per_token": 4e-07, - "litellm_provider": "vertex_ai-mistral_models", - "max_input_tokens": 128000, - "max_output_tokens": 8191, - "max_tokens": 8191, - "mode": "chat", - "output_cost_per_token": 2e-06, - "supports_function_calling": true, - "supports_tool_choice": true - }, "vertex_ai/mistral-nemo@2407": { "input_cost_per_token": 3e-06, "litellm_provider": "vertex_ai-mistral_models", @@ -35268,15 +31012,6 @@ "supports_function_calling": true, "supports_tool_choice": true }, - "vertex_ai/mistral-ocr-2505": { - "litellm_provider": "vertex_ai", - "mode": "ocr", - "ocr_cost_per_page": 0.0005, - "supported_endpoints": [ - "/v1/ocr" - ], - "source": "https://cloud.google.com/generative-ai-app-builder/pricing" - }, "vertex_ai/mistral-small-2503": { "input_cost_per_token": 1e-06, "litellm_provider": "vertex_ai-mistral_models", @@ -35300,62 +31035,22 @@ "supports_function_calling": true, "supports_tool_choice": true }, - "vertex_ai/mistralai/codestral-2": { + "vertex_ai/mistral-ocr-2505": { + "litellm_provider": "vertex_ai", + "mode": "ocr", + "ocr_cost_per_page": 0.0005, + "supported_endpoints": [ + "/v1/ocr" + ], + "source": "https://cloud.google.com/generative-ai-app-builder/pricing" + }, + "vertex_ai/deepseek-ai/deepseek-ocr-maas": { + "litellm_provider": "vertex_ai", + "mode": "ocr", "input_cost_per_token": 3e-07, - "litellm_provider": "vertex_ai-mistral_models", - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 9e-07, - "supports_function_calling": true, - "supports_tool_choice": true - }, - "vertex_ai/mistralai/codestral-2@001": { - "input_cost_per_token": 3e-07, - "litellm_provider": "vertex_ai-mistral_models", - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 9e-07, - "supports_function_calling": true, - "supports_tool_choice": true - }, - "vertex_ai/mistralai/mistral-medium-3": { - "input_cost_per_token": 4e-07, - "litellm_provider": "vertex_ai-mistral_models", - "max_input_tokens": 128000, - "max_output_tokens": 8191, - "max_tokens": 8191, - "mode": "chat", - "output_cost_per_token": 2e-06, - "supports_function_calling": true, - "supports_tool_choice": true - }, - "vertex_ai/mistralai/mistral-medium-3@001": { - "input_cost_per_token": 4e-07, - "litellm_provider": "vertex_ai-mistral_models", - "max_input_tokens": 128000, - "max_output_tokens": 8191, - "max_tokens": 8191, - "mode": "chat", - "output_cost_per_token": 2e-06, - "supports_function_calling": true, - "supports_tool_choice": true - }, - "vertex_ai/moonshotai/kimi-k2-thinking-maas": { - "input_cost_per_token": 6e-07, - "litellm_provider": "vertex_ai-moonshot_models", - "max_input_tokens": 256000, - "max_output_tokens": 256000, - "max_tokens": 256000, - "mode": "chat", - "output_cost_per_token": 2.5e-06, - "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#partner-models", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_web_search": true + "output_cost_per_token": 1.2e-06, + "ocr_cost_per_page": 0.0003, + "source": "https://cloud.google.com/vertex-ai/pricing" }, "vertex_ai/openai/gpt-oss-120b-maas": { "input_cost_per_token": 1.5e-07, @@ -35439,11 +31134,6 @@ "supports_function_calling": true, "supports_tool_choice": true }, - "vertex_ai/search_api": { - "input_cost_per_query": 0.0015, - "litellm_provider": "vertex_ai", - "mode": "vector_store" - }, "vertex_ai/veo-2.0-generate-001": { "litellm_provider": "vertex_ai-video-models", "max_input_tokens": 1024, @@ -35486,12 +31176,12 @@ "video" ] }, - "vertex_ai/veo-3.1-fast-generate-001": { + "vertex_ai/veo-3.1-generate-preview": { "litellm_provider": "vertex_ai-video-models", "max_input_tokens": 1024, "max_tokens": 1024, "mode": "video_generation", - "output_cost_per_second": 0.15, + "output_cost_per_second": 0.4, "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/veo", "supported_modalities": [ "text" @@ -35528,12 +31218,12 @@ "video" ] }, - "vertex_ai/veo-3.1-generate-preview": { + "vertex_ai/veo-3.1-fast-generate-001": { "litellm_provider": "vertex_ai-video-models", "max_input_tokens": 1024, "max_tokens": 1024, "mode": "video_generation", - "output_cost_per_second": 0.4, + "output_cost_per_second": 0.15, "source": "https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/veo", "supported_modalities": [ "text" @@ -35542,34 +31232,6 @@ "video" ] }, - "vertex_ai/zai-org/glm-4.7-maas": { - "input_cost_per_token": 6e-07, - "litellm_provider": "vertex_ai-zai_models", - "max_input_tokens": 200000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 2.2e-06, - "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#partner-models", - "supports_function_calling": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, - "vertex_ai/zai-org/glm-5-maas": { - "cache_read_input_token_cost": 1e-07, - "input_cost_per_token": 1e-06, - "litellm_provider": "vertex_ai-zai_models", - "max_input_tokens": 200000, - "max_output_tokens": 128000, - "max_tokens": 128000, - "mode": "chat", - "output_cost_per_token": 3.2e-06, - "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing#glm-models", - "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_tool_choice": true - }, "voyage/rerank-2": { "input_cost_per_token": 5e-08, "litellm_provider": "voyage", @@ -35730,16 +31392,34 @@ "mode": "embedding", "output_cost_per_token": 0.0 }, - "wandb/Qwen/Qwen3-235B-A22B-Instruct-2507": { - "max_tokens": 262144, - "max_input_tokens": 262144, - "max_output_tokens": 262144, - "input_cost_per_token": 0.01, - "output_cost_per_token": 0.01, + "wandb/openai/gpt-oss-120b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 0.015, + "output_cost_per_token": 0.06, "litellm_provider": "wandb", "mode": "chat" }, - "wandb/Qwen/Qwen3-235B-A22B-Thinking-2507": { + "wandb/openai/gpt-oss-20b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 0.005, + "output_cost_per_token": 0.02, + "litellm_provider": "wandb", + "mode": "chat" + }, + "wandb/zai-org/GLM-4.5": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 0.055, + "output_cost_per_token": 0.2, + "litellm_provider": "wandb", + "mode": "chat" + }, + "wandb/Qwen/Qwen3-235B-A22B-Instruct-2507": { "max_tokens": 262144, "max_input_tokens": 262144, "max_output_tokens": 262144, @@ -35757,6 +31437,42 @@ "litellm_provider": "wandb", "mode": "chat" }, + "wandb/Qwen/Qwen3-235B-A22B-Thinking-2507": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 0.01, + "output_cost_per_token": 0.01, + "litellm_provider": "wandb", + "mode": "chat" + }, + "wandb/moonshotai/Kimi-K2-Instruct": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 6e-07, + "output_cost_per_token": 2.5e-06, + "litellm_provider": "wandb", + "mode": "chat" + }, + "wandb/meta-llama/Llama-3.1-8B-Instruct": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 0.022, + "output_cost_per_token": 0.022, + "litellm_provider": "wandb", + "mode": "chat" + }, + "wandb/deepseek-ai/DeepSeek-V3.1": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 0.055, + "output_cost_per_token": 0.165, + "litellm_provider": "wandb", + "mode": "chat" + }, "wandb/deepseek-ai/DeepSeek-R1-0528": { "max_tokens": 161000, "max_input_tokens": 161000, @@ -35775,24 +31491,6 @@ "litellm_provider": "wandb", "mode": "chat" }, - "wandb/deepseek-ai/DeepSeek-V3.1": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 0.055, - "output_cost_per_token": 0.165, - "litellm_provider": "wandb", - "mode": "chat" - }, - "wandb/meta-llama/Llama-3.1-8B-Instruct": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 0.022, - "output_cost_per_token": 0.022, - "litellm_provider": "wandb", - "mode": "chat" - }, "wandb/meta-llama/Llama-3.3-70B-Instruct": { "max_tokens": 128000, "max_input_tokens": 128000, @@ -35820,41 +31518,41 @@ "litellm_provider": "wandb", "mode": "chat" }, - "wandb/moonshotai/Kimi-K2-Instruct": { - "max_tokens": 128000, - "max_input_tokens": 128000, - "max_output_tokens": 128000, - "input_cost_per_token": 6e-07, - "output_cost_per_token": 2.5e-06, - "litellm_provider": "wandb", - "mode": "chat" + "watsonx/ibm/granite-3-8b-instruct": { + "input_cost_per_token": 2e-07, + "litellm_provider": "watsonx", + "max_input_tokens": 8192, + "max_output_tokens": 1024, + "max_tokens": 1024, + "mode": "chat", + "output_cost_per_token": 2e-07, + "supports_audio_input": false, + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": false, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": false }, - "wandb/openai/gpt-oss-120b": { - "max_tokens": 131072, + "watsonx/mistralai/mistral-large": { + "input_cost_per_token": 3e-06, + "litellm_provider": "watsonx", "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 0.015, - "output_cost_per_token": 0.06, - "litellm_provider": "wandb", - "mode": "chat" - }, - "wandb/openai/gpt-oss-20b": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 0.005, - "output_cost_per_token": 0.02, - "litellm_provider": "wandb", - "mode": "chat" - }, - "wandb/zai-org/GLM-4.5": { - "max_tokens": 131072, - "max_input_tokens": 131072, - "max_output_tokens": 131072, - "input_cost_per_token": 0.055, - "output_cost_per_token": 0.2, - "litellm_provider": "wandb", - "mode": "chat" + "max_output_tokens": 16384, + "max_tokens": 16384, + "mode": "chat", + "output_cost_per_token": 1e-05, + "supports_audio_input": false, + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": false, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": false }, "watsonx/bigscience/mt0-xxl-13b": { "max_tokens": 8192, @@ -35928,24 +31626,6 @@ "supports_parallel_function_calling": true, "supports_vision": false }, - "watsonx/ibm/granite-3-8b-instruct": { - "input_cost_per_token": 2e-07, - "litellm_provider": "watsonx", - "max_input_tokens": 8192, - "max_output_tokens": 1024, - "max_tokens": 1024, - "mode": "chat", - "output_cost_per_token": 2e-07, - "supports_audio_input": false, - "supports_audio_output": false, - "supports_function_calling": true, - "supports_parallel_function_calling": false, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": false - }, "watsonx/ibm/granite-4-h-small": { "max_tokens": 20480, "max_input_tokens": 20480, @@ -36114,24 +31794,6 @@ "supports_parallel_function_calling": false, "supports_vision": true }, - "watsonx/mistralai/mistral-large": { - "input_cost_per_token": 3e-06, - "litellm_provider": "watsonx", - "max_input_tokens": 131072, - "max_output_tokens": 16384, - "max_tokens": 16384, - "mode": "chat", - "output_cost_per_token": 1e-05, - "supports_audio_input": false, - "supports_audio_output": false, - "supports_function_calling": true, - "supports_parallel_function_calling": false, - "supports_prompt_caching": true, - "supports_response_schema": true, - "supports_system_messages": true, - "supports_tool_choice": true, - "supports_vision": false - }, "watsonx/mistralai/mistral-medium-2505": { "max_tokens": 128000, "max_input_tokens": 128000, @@ -36222,28 +31884,6 @@ "/v1/audio/transcriptions" ] }, - "writer.palmyra-x4-v1:0": { - "input_cost_per_token": 2.5e-06, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 128000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 1e-05, - "supports_function_calling": true, - "supports_pdf_input": true - }, - "writer.palmyra-x5-v1:0": { - "input_cost_per_token": 6e-07, - "litellm_provider": "bedrock_converse", - "max_input_tokens": 1000000, - "max_output_tokens": 8192, - "max_tokens": 8192, - "mode": "chat", - "output_cost_per_token": 6e-06, - "supports_function_calling": true, - "supports_pdf_input": true - }, "xai/grok-2": { "input_cost_per_token": 2e-06, "litellm_provider": "xai", @@ -36509,6 +32149,38 @@ "supports_tool_choice": true, "supports_web_search": true }, + "xai/grok-4-fast-reasoning": { + "litellm_provider": "xai", + "max_input_tokens": 2000000.0, + "max_output_tokens": 2000000.0, + "max_tokens": 2000000.0, + "mode": "chat", + "input_cost_per_token": 2e-07, + "input_cost_per_token_above_128k_tokens": 4e-07, + "output_cost_per_token": 5e-07, + "output_cost_per_token_above_128k_tokens": 1e-06, + "cache_read_input_token_cost": 5e-08, + "source": "https://docs.x.ai/docs/models", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_web_search": true + }, + "xai/grok-4-fast-non-reasoning": { + "litellm_provider": "xai", + "max_input_tokens": 2000000.0, + "max_output_tokens": 2000000.0, + "cache_read_input_token_cost": 5e-08, + "max_tokens": 2000000.0, + "mode": "chat", + "input_cost_per_token": 2e-07, + "input_cost_per_token_above_128k_tokens": 4e-07, + "output_cost_per_token": 5e-07, + "output_cost_per_token_above_128k_tokens": 1e-06, + "source": "https://docs.x.ai/docs/models", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_web_search": true + }, "xai/grok-4-0709": { "input_cost_per_token": 3e-06, "input_cost_per_token_above_128k_tokens": 6e-06, @@ -36524,6 +32196,21 @@ "supports_tool_choice": true, "supports_web_search": true }, + "xai/grok-4-latest": { + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_128k_tokens": 6e-06, + "litellm_provider": "xai", + "max_input_tokens": 256000, + "max_output_tokens": 256000, + "max_tokens": 256000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "output_cost_per_token_above_128k_tokens": 3e-05, + "source": "https://docs.x.ai/docs/models", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_web_search": true + }, "xai/grok-4-1-fast": { "cache_read_input_token_cost": 5e-08, "input_cost_per_token": 2e-07, @@ -36544,6 +32231,46 @@ "supports_vision": true, "supports_web_search": true }, + "xai/grok-4-1-fast-reasoning": { + "cache_read_input_token_cost": 5e-08, + "input_cost_per_token": 2e-07, + "input_cost_per_token_above_128k_tokens": 4e-07, + "litellm_provider": "xai", + "max_input_tokens": 2000000.0, + "max_output_tokens": 2000000.0, + "max_tokens": 2000000.0, + "mode": "chat", + "output_cost_per_token": 5e-07, + "output_cost_per_token_above_128k_tokens": 1e-06, + "source": "https://docs.x.ai/docs/models/grok-4-1-fast-reasoning", + "supports_audio_input": true, + "supports_function_calling": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true + }, + "xai/grok-4-1-fast-reasoning-latest": { + "cache_read_input_token_cost": 5e-08, + "input_cost_per_token": 2e-07, + "input_cost_per_token_above_128k_tokens": 4e-07, + "litellm_provider": "xai", + "max_input_tokens": 2000000.0, + "max_output_tokens": 2000000.0, + "max_tokens": 2000000.0, + "mode": "chat", + "output_cost_per_token": 5e-07, + "output_cost_per_token_above_128k_tokens": 1e-06, + "source": "https://docs.x.ai/docs/models/grok-4-1-fast-reasoning", + "supports_audio_input": true, + "supports_function_calling": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true + }, "xai/grok-4-1-fast-non-reasoning": { "cache_read_input_token_cost": 5e-08, "input_cost_per_token": 2e-07, @@ -36582,93 +32309,6 @@ "supports_vision": true, "supports_web_search": true }, - "xai/grok-4-1-fast-reasoning": { - "cache_read_input_token_cost": 5e-08, - "input_cost_per_token": 2e-07, - "input_cost_per_token_above_128k_tokens": 4e-07, - "litellm_provider": "xai", - "max_input_tokens": 2000000.0, - "max_output_tokens": 2000000.0, - "max_tokens": 2000000.0, - "mode": "chat", - "output_cost_per_token": 5e-07, - "output_cost_per_token_above_128k_tokens": 1e-06, - "source": "https://docs.x.ai/docs/models/grok-4-1-fast-reasoning", - "supports_audio_input": true, - "supports_function_calling": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true - }, - "xai/grok-4-1-fast-reasoning-latest": { - "cache_read_input_token_cost": 5e-08, - "input_cost_per_token": 2e-07, - "input_cost_per_token_above_128k_tokens": 4e-07, - "litellm_provider": "xai", - "max_input_tokens": 2000000.0, - "max_output_tokens": 2000000.0, - "max_tokens": 2000000.0, - "mode": "chat", - "output_cost_per_token": 5e-07, - "output_cost_per_token_above_128k_tokens": 1e-06, - "source": "https://docs.x.ai/docs/models/grok-4-1-fast-reasoning", - "supports_audio_input": true, - "supports_function_calling": true, - "supports_reasoning": true, - "supports_response_schema": true, - "supports_tool_choice": true, - "supports_vision": true, - "supports_web_search": true - }, - "xai/grok-4-fast-non-reasoning": { - "litellm_provider": "xai", - "max_input_tokens": 2000000.0, - "max_output_tokens": 2000000.0, - "cache_read_input_token_cost": 5e-08, - "max_tokens": 2000000.0, - "mode": "chat", - "input_cost_per_token": 2e-07, - "input_cost_per_token_above_128k_tokens": 4e-07, - "output_cost_per_token": 5e-07, - "output_cost_per_token_above_128k_tokens": 1e-06, - "source": "https://docs.x.ai/docs/models", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_web_search": true - }, - "xai/grok-4-fast-reasoning": { - "litellm_provider": "xai", - "max_input_tokens": 2000000.0, - "max_output_tokens": 2000000.0, - "max_tokens": 2000000.0, - "mode": "chat", - "input_cost_per_token": 2e-07, - "input_cost_per_token_above_128k_tokens": 4e-07, - "output_cost_per_token": 5e-07, - "output_cost_per_token_above_128k_tokens": 1e-06, - "cache_read_input_token_cost": 5e-08, - "source": "https://docs.x.ai/docs/models", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_web_search": true - }, - "xai/grok-4-latest": { - "input_cost_per_token": 3e-06, - "input_cost_per_token_above_128k_tokens": 6e-06, - "litellm_provider": "xai", - "max_input_tokens": 256000, - "max_output_tokens": 256000, - "max_tokens": 256000, - "mode": "chat", - "output_cost_per_token": 1.5e-05, - "output_cost_per_token_above_128k_tokens": 3e-05, - "source": "https://docs.x.ai/docs/models", - "supports_function_calling": true, - "supports_tool_choice": true, - "supports_web_search": true - }, "xai/grok-beta": { "input_cost_per_token": 5e-06, "litellm_provider": "xai", @@ -36766,9 +32406,69 @@ "supports_tool_choice": true, "source": "https://aws.amazon.com/bedrock/pricing/" }, - "zai/glm-4-32b-0414-128k": { - "input_cost_per_token": 1e-07, - "output_cost_per_token": 1e-07, + "zai/glm-5": { + "cache_creation_input_token_cost": 0, + "cache_read_input_token_cost": 2e-07, + "input_cost_per_token": 1e-06, + "output_cost_per_token": 3.2e-06, + "litellm_provider": "zai", + "max_input_tokens": 200000, + "max_output_tokens": 128000, + "mode": "chat", + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "source": "https://docs.z.ai/guides/overview/pricing" + }, + "zai/glm-5-code": { + "cache_creation_input_token_cost": 0, + "cache_read_input_token_cost": 3e-07, + "input_cost_per_token": 1.2e-06, + "output_cost_per_token": 5e-06, + "litellm_provider": "zai", + "max_input_tokens": 200000, + "max_output_tokens": 128000, + "mode": "chat", + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "source": "https://docs.z.ai/guides/overview/pricing" + }, + "zai/glm-4.7": { + "cache_creation_input_token_cost": 0, + "cache_read_input_token_cost": 1.1e-07, + "input_cost_per_token": 6e-07, + "output_cost_per_token": 2.2e-06, + "litellm_provider": "zai", + "max_input_tokens": 200000, + "max_output_tokens": 128000, + "mode": "chat", + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "source": "https://docs.z.ai/guides/overview/pricing" + }, + "zai/glm-4.6": { + "cache_creation_input_token_cost": 0, + "cache_read_input_token_cost": 1.1e-07, + "input_cost_per_token": 6e-07, + "output_cost_per_token": 2.2e-06, + "litellm_provider": "zai", + "max_input_tokens": 200000, + "max_output_tokens": 128000, + "mode": "chat", + "supports_function_calling": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_tool_choice": true, + "source": "https://docs.z.ai/guides/overview/pricing" + }, + "zai/glm-4.5": { + "input_cost_per_token": 6e-07, + "output_cost_per_token": 2.2e-06, "litellm_provider": "zai", "max_input_tokens": 128000, "max_output_tokens": 32000, @@ -36777,9 +32477,21 @@ "supports_tool_choice": true, "source": "https://docs.z.ai/guides/overview/pricing" }, - "zai/glm-4.5": { + "zai/glm-4.5v": { "input_cost_per_token": 6e-07, - "output_cost_per_token": 2.2e-06, + "output_cost_per_token": 1.8e-06, + "litellm_provider": "zai", + "max_input_tokens": 128000, + "max_output_tokens": 32000, + "mode": "chat", + "supports_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true, + "source": "https://docs.z.ai/guides/overview/pricing" + }, + "zai/glm-4.5-x": { + "input_cost_per_token": 2.2e-06, + "output_cost_per_token": 8.9e-06, "litellm_provider": "zai", "max_input_tokens": 128000, "max_output_tokens": 32000, @@ -36810,6 +32522,17 @@ "supports_tool_choice": true, "source": "https://docs.z.ai/guides/overview/pricing" }, + "zai/glm-4-32b-0414-128k": { + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "zai", + "max_input_tokens": 128000, + "max_output_tokens": 32000, + "mode": "chat", + "supports_function_calling": true, + "supports_tool_choice": true, + "source": "https://docs.z.ai/guides/overview/pricing" + }, "zai/glm-4.5-flash": { "input_cost_per_token": 0, "output_cost_per_token": 0, @@ -36821,87 +32544,4401 @@ "supports_tool_choice": true, "source": "https://docs.z.ai/guides/overview/pricing" }, - "zai/glm-4.5-x": { - "input_cost_per_token": 2.2e-06, - "output_cost_per_token": 8.9e-06, - "litellm_provider": "zai", - "max_input_tokens": 128000, - "max_output_tokens": 32000, - "mode": "chat", - "supports_function_calling": true, - "supports_tool_choice": true, - "source": "https://docs.z.ai/guides/overview/pricing" + "vertex_ai/search_api": { + "input_cost_per_query": 0.0015, + "litellm_provider": "vertex_ai", + "mode": "vector_store" }, - "zai/glm-4.5v": { - "input_cost_per_token": 6e-07, + "openai/container": { + "code_interpreter_cost_per_session": 0.03, + "litellm_provider": "openai", + "mode": "chat" + }, + "openai/sora-2": { + "litellm_provider": "openai", + "mode": "video_generation", + "output_cost_per_video_per_second": 0.1, + "source": "https://platform.openai.com/docs/api-reference/videos", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "video" + ], + "supported_resolutions": [ + "720x1280", + "1280x720" + ] + }, + "openai/sora-2-pro": { + "litellm_provider": "openai", + "mode": "video_generation", + "output_cost_per_video_per_second": 0.3, + "source": "https://platform.openai.com/docs/api-reference/videos", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "video" + ], + "supported_resolutions": [ + "720x1280", + "1280x720" + ] + }, + "openai/sora-2-pro-high-res": { + "litellm_provider": "openai", + "mode": "video_generation", + "output_cost_per_video_per_second": 0.5, + "source": "https://platform.openai.com/docs/api-reference/videos", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "video" + ], + "supported_resolutions": [ + "1024x1792", + "1792x1024" + ] + }, + "azure/sora-2": { + "litellm_provider": "azure", + "mode": "video_generation", + "output_cost_per_video_per_second": 0.1, + "source": "https://azure.microsoft.com/en-us/products/ai-services/video-generation", + "supported_modalities": [ + "text" + ], + "supported_output_modalities": [ + "video" + ], + "supported_resolutions": [ + "720x1280", + "1280x720" + ] + }, + "azure/sora-2-pro": { + "litellm_provider": "azure", + "mode": "video_generation", + "output_cost_per_video_per_second": 0.3, + "source": "https://azure.microsoft.com/en-us/products/ai-services/video-generation", + "supported_modalities": [ + "text" + ], + "supported_output_modalities": [ + "video" + ], + "supported_resolutions": [ + "720x1280", + "1280x720" + ] + }, + "azure/sora-2-pro-high-res": { + "litellm_provider": "azure", + "mode": "video_generation", + "output_cost_per_video_per_second": 0.5, + "source": "https://azure.microsoft.com/en-us/products/ai-services/video-generation", + "supported_modalities": [ + "text" + ], + "supported_output_modalities": [ + "video" + ], + "supported_resolutions": [ + "1024x1792", + "1792x1024" + ] + }, + "runwayml/gen4_turbo": { + "litellm_provider": "runwayml", + "mode": "video_generation", + "output_cost_per_video_per_second": 0.05, + "source": "https://docs.dev.runwayml.com/guides/pricing/", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "video" + ], + "supported_resolutions": [ + "1280x720", + "720x1280" + ], + "metadata": { + "comment": "5 credits per second @ $0.01 per credit = $0.05 per second" + } + }, + "runwayml/gen4_aleph": { + "litellm_provider": "runwayml", + "mode": "video_generation", + "output_cost_per_video_per_second": 0.15, + "source": "https://docs.dev.runwayml.com/guides/pricing/", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "video" + ], + "supported_resolutions": [ + "1280x720", + "720x1280" + ], + "metadata": { + "comment": "15 credits per second @ $0.01 per credit = $0.15 per second" + } + }, + "runwayml/gen3a_turbo": { + "litellm_provider": "runwayml", + "mode": "video_generation", + "output_cost_per_video_per_second": 0.05, + "source": "https://docs.dev.runwayml.com/guides/pricing/", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "video" + ], + "supported_resolutions": [ + "1280x720", + "720x1280" + ], + "metadata": { + "comment": "5 credits per second @ $0.01 per credit = $0.05 per second" + } + }, + "runwayml/gen4_image": { + "litellm_provider": "runwayml", + "mode": "image_generation", + "input_cost_per_image": 0.05, + "output_cost_per_image": 0.05, + "source": "https://docs.dev.runwayml.com/guides/pricing/", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "image" + ], + "supported_resolutions": [ + "1280x720", + "1920x1080" + ], + "metadata": { + "comment": "5 credits per 720p image or 8 credits per 1080p image @ $0.01 per credit. Using 5 credits ($0.05) as base cost" + } + }, + "runwayml/gen4_image_turbo": { + "litellm_provider": "runwayml", + "mode": "image_generation", + "input_cost_per_image": 0.02, + "output_cost_per_image": 0.02, + "source": "https://docs.dev.runwayml.com/guides/pricing/", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "image" + ], + "supported_resolutions": [ + "1280x720", + "1920x1080" + ], + "metadata": { + "comment": "2 credits per image (any resolution) @ $0.01 per credit = $0.02 per image" + } + }, + "runwayml/eleven_multilingual_v2": { + "litellm_provider": "runwayml", + "mode": "audio_speech", + "input_cost_per_character": 3e-07, + "source": "https://docs.dev.runwayml.com/guides/pricing/", + "metadata": { + "comment": "Estimated cost based on standard TTS pricing. RunwayML uses ElevenLabs models." + } + }, + "fireworks_ai/accounts/fireworks/models/qwen3-coder-480b-a35b-instruct": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 4.5e-07, "output_cost_per_token": 1.8e-06, - "litellm_provider": "zai", - "max_input_tokens": 128000, - "max_output_tokens": 32000, + "litellm_provider": "fireworks_ai", "mode": "chat", + "supports_reasoning": true + }, + "fireworks_ai/accounts/fireworks/models/flux-kontext-pro": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 4e-08, + "output_cost_per_token": 4e-08, + "litellm_provider": "fireworks_ai", + "mode": "image_generation" + }, + "fireworks_ai/accounts/fireworks/models/SSD-1B": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1.3e-10, + "output_cost_per_token": 1.3e-10, + "litellm_provider": "fireworks_ai", + "mode": "image_generation" + }, + "fireworks_ai/accounts/fireworks/models/chronos-hermes-13b-v2": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-llama-13b": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-llama-13b-instruct": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-llama-13b-python": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-llama-34b": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-llama-34b-instruct": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-llama-34b-python": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-llama-70b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-llama-70b-instruct": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-llama-70b-python": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-llama-7b": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-llama-7b-instruct": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-llama-7b-python": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/code-qwen-1p5-7b": { + "max_tokens": 65536, + "max_input_tokens": 65536, + "max_output_tokens": 65536, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/codegemma-2b": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/codegemma-7b": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/cogito-671b-v2-p1": { + "max_tokens": 163840, + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "input_cost_per_token": 1.2e-06, + "output_cost_per_token": 1.2e-06, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/cogito-v1-preview-llama-3b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/cogito-v1-preview-llama-70b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/cogito-v1-preview-llama-8b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/cogito-v1-preview-qwen-14b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/cogito-v1-preview-qwen-32b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/flux-kontext-max": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 8e-08, + "output_cost_per_token": 8e-08, + "litellm_provider": "fireworks_ai", + "mode": "image_generation" + }, + "fireworks_ai/accounts/fireworks/models/dbrx-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1.2e-06, + "output_cost_per_token": 1.2e-06, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-coder-1b-base": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-coder-33b-instruct": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-coder-7b-base": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-coder-7b-base-v1p5": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-coder-7b-instruct-v1p5": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-coder-v2-lite-base": { + "max_tokens": 163840, + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-coder-v2-lite-instruct": { + "max_tokens": 163840, + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-prover-v2": { + "max_tokens": 163840, + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "input_cost_per_token": 1.2e-06, + "output_cost_per_token": 1.2e-06, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-r1-0528-distill-qwen3-8b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-r1-distill-llama-70b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-r1-distill-llama-8b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-r1-distill-qwen-14b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-r1-distill-qwen-1p5b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-r1-distill-qwen-32b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-r1-distill-qwen-7b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-v2-lite-chat": { + "max_tokens": 163840, + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/deepseek-v2p5": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1.2e-06, + "output_cost_per_token": 1.2e-06, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/devstral-small-2505": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/dobby-mini-unhinged-plus-llama-3-1-8b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/dobby-unhinged-llama-3-3-70b-new": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/dolphin-2-9-2-qwen2-72b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/dolphin-2p6-mixtral-8x7b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/ernie-4p5-21b-a3b-pt": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/ernie-4p5-300b-a47b-pt": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/fare-20b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/firefunction-v1": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/firellava-13b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/firesearch-ocr-v6": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/fireworks-asr-large": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 0.0, + "output_cost_per_token": 0.0, + "litellm_provider": "fireworks_ai", + "mode": "audio_transcription" + }, + "fireworks_ai/accounts/fireworks/models/fireworks-asr-v2": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 0.0, + "output_cost_per_token": 0.0, + "litellm_provider": "fireworks_ai", + "mode": "audio_transcription" + }, + "fireworks_ai/accounts/fireworks/models/flux-1-dev": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/flux-1-dev-controlnet-union": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1e-09, + "output_cost_per_token": 1e-09, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/flux-1-dev-fp8": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 5e-10, + "output_cost_per_token": 5e-10, + "litellm_provider": "fireworks_ai", + "mode": "image_generation" + }, + "fireworks_ai/accounts/fireworks/models/flux-1-schnell": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/flux-1-schnell-fp8": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 3.5e-10, + "output_cost_per_token": 3.5e-10, + "litellm_provider": "fireworks_ai", + "mode": "image_generation" + }, + "fireworks_ai/accounts/fireworks/models/gemma-2b-it": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/gemma-3-27b-it": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/gemma-7b": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/gemma-7b-it": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/gemma2-9b-it": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/glm-4p5v": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1.2e-06, + "output_cost_per_token": 1.2e-06, + "litellm_provider": "fireworks_ai", + "mode": "chat", + "supports_reasoning": true + }, + "fireworks_ai/accounts/fireworks/models/gpt-oss-safeguard-120b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1.2e-06, + "output_cost_per_token": 1.2e-06, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/gpt-oss-safeguard-20b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/hermes-2-pro-mistral-7b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/internvl3-38b": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/internvl3-78b": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/internvl3-8b": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/japanese-stable-diffusion-xl": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1.3e-10, + "output_cost_per_token": 1.3e-10, + "litellm_provider": "fireworks_ai", + "mode": "image_generation" + }, + "fireworks_ai/accounts/fireworks/models/kat-coder": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/kat-dev-32b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/kat-dev-72b-exp": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-guard-2-8b": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-guard-3-1b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-guard-3-8b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v2-13b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v2-13b-chat": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v2-70b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v2-70b-chat": { + "max_tokens": 2048, + "max_input_tokens": 2048, + "max_output_tokens": 2048, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v2-7b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v2-7b-chat": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v3-70b-instruct": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v3-70b-instruct-hf": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v3-8b": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v3-8b-instruct-hf": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v3p1-405b-instruct-long": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v3p1-70b-instruct": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v3p1-70b-instruct-1b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v3p1-nemotron-70b-instruct": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v3p2-1b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v3p2-3b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llama-v3p3-70b-instruct": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llamaguard-7b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/llava-yi-34b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/minimax-m1-80k": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/minimax-m2": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 3e-07, + "output_cost_per_token": 1.2e-06, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/ministral-3-14b-instruct-2512": { + "max_tokens": 256000, + "max_input_tokens": 256000, + "max_output_tokens": 256000, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/ministral-3-3b-instruct-2512": { + "max_tokens": 256000, + "max_input_tokens": 256000, + "max_output_tokens": 256000, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/ministral-3-8b-instruct-2512": { + "max_tokens": 256000, + "max_input_tokens": 256000, + "max_output_tokens": 256000, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mistral-7b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mistral-7b-instruct-4k": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mistral-7b-instruct-v0p2": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mistral-7b-instruct-v3": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mistral-7b-v0p2": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mistral-large-3-fp8": { + "max_tokens": 256000, + "max_input_tokens": 256000, + "max_output_tokens": 256000, + "input_cost_per_token": 1.2e-06, + "output_cost_per_token": 1.2e-06, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mistral-nemo-base-2407": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mistral-nemo-instruct-2407": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mistral-small-24b-instruct-2501": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mixtral-8x22b": { + "max_tokens": 65536, + "max_input_tokens": 65536, + "max_output_tokens": 65536, + "input_cost_per_token": 1.2e-06, + "output_cost_per_token": 1.2e-06, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mixtral-8x22b-instruct": { + "max_tokens": 65536, + "max_input_tokens": 65536, + "max_output_tokens": 65536, + "input_cost_per_token": 1.2e-06, + "output_cost_per_token": 1.2e-06, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mixtral-8x7b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mixtral-8x7b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mixtral-8x7b-instruct-hf": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/mythomax-l2-13b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/nemotron-nano-v2-12b-vl": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/nous-capybara-7b-v1p9": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/nous-hermes-2-mixtral-8x7b-dpo": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/nous-hermes-2-yi-34b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/nous-hermes-llama2-13b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/nous-hermes-llama2-70b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/nous-hermes-llama2-7b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/nvidia-nemotron-nano-12b-v2": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/nvidia-nemotron-nano-9b-v2": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/openchat-3p5-0106-7b": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/openhermes-2-mistral-7b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/openhermes-2p5-mistral-7b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/openorca-7b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/phi-2-3b": { + "max_tokens": 2048, + "max_input_tokens": 2048, + "max_output_tokens": 2048, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/phi-3-mini-128k-instruct": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/phi-3-vision-128k-instruct": { + "max_tokens": 32064, + "max_input_tokens": 32064, + "max_output_tokens": 32064, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/phind-code-llama-34b-python-v1": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/phind-code-llama-34b-v1": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/phind-code-llama-34b-v2": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/playground-v2-1024px-aesthetic": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1.3e-10, + "output_cost_per_token": 1.3e-10, + "litellm_provider": "fireworks_ai", + "mode": "image_generation" + }, + "fireworks_ai/accounts/fireworks/models/playground-v2-5-1024px-aesthetic": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1.3e-10, + "output_cost_per_token": 1.3e-10, + "litellm_provider": "fireworks_ai", + "mode": "image_generation" + }, + "fireworks_ai/accounts/fireworks/models/pythia-12b": { + "max_tokens": 2048, + "max_input_tokens": 2048, + "max_output_tokens": 2048, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen-qwq-32b-preview": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen-v2p5-14b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen-v2p5-7b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen1p5-72b-chat": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2-7b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2-vl-2b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2-vl-72b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2-vl-7b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-0p5b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-14b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-1p5b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-32b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-32b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-72b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-72b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-7b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-0p5b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-0p5b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-14b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-14b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-1p5b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-1p5b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-32b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-32b-instruct-128k": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-32b-instruct-32k-rope": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-32b-instruct-64k": { + "max_tokens": 65536, + "max_input_tokens": 65536, + "max_output_tokens": 65536, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-3b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-3b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-7b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-coder-7b-instruct": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-math-72b-instruct": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-vl-32b-instruct": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-vl-3b-instruct": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-vl-72b-instruct": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen2p5-vl-7b-instruct": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-0p6b": { + "max_tokens": 40960, + "max_input_tokens": 40960, + "max_output_tokens": 40960, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-14b": { + "max_tokens": 40960, + "max_input_tokens": 40960, + "max_output_tokens": 40960, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-1p7b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-1p7b-fp8-draft": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-1p7b-fp8-draft-131072": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-1p7b-fp8-draft-40960": { + "max_tokens": 40960, + "max_input_tokens": 40960, + "max_output_tokens": 40960, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-235b-a22b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 2.2e-07, + "output_cost_per_token": 8.8e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-235b-a22b-instruct-2507": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 2.2e-07, + "output_cost_per_token": 8.8e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-235b-a22b-thinking-2507": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 2.2e-07, + "output_cost_per_token": 8.8e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-30b-a3b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 1.5e-07, + "output_cost_per_token": 6e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-30b-a3b-instruct-2507": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 5e-07, + "output_cost_per_token": 5e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-30b-a3b-thinking-2507": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-32b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat", + "supports_reasoning": true + }, + "fireworks_ai/accounts/fireworks/models/qwen3-4b": { + "max_tokens": 40960, + "max_input_tokens": 40960, + "max_output_tokens": 40960, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-4b-instruct-2507": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-8b": { + "max_tokens": 40960, + "max_input_tokens": 40960, + "max_output_tokens": 40960, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat", + "supports_reasoning": true + }, + "fireworks_ai/accounts/fireworks/models/qwen3-coder-30b-a3b-instruct": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 1.5e-07, + "output_cost_per_token": 6e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-coder-480b-instruct-bf16": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-embedding-0p6b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 0.0, + "output_cost_per_token": 0.0, + "litellm_provider": "fireworks_ai", + "mode": "embedding" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-embedding-4b": { + "max_tokens": 40960, + "max_input_tokens": 40960, + "max_output_tokens": 40960, + "input_cost_per_token": 0.0, + "output_cost_per_token": 0.0, + "litellm_provider": "fireworks_ai", + "mode": "embedding" + }, + "fireworks_ai/accounts/fireworks/models/": { + "max_tokens": 40960, + "max_input_tokens": 40960, + "max_output_tokens": 40960, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 0.0, + "litellm_provider": "fireworks_ai", + "mode": "embedding" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-next-80b-a3b-instruct": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-next-80b-a3b-thinking": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-reranker-0p6b": { + "max_tokens": 40960, + "max_input_tokens": 40960, + "max_output_tokens": 40960, + "input_cost_per_token": 0.0, + "output_cost_per_token": 0.0, + "litellm_provider": "fireworks_ai", + "mode": "rerank" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-reranker-4b": { + "max_tokens": 40960, + "max_input_tokens": 40960, + "max_output_tokens": 40960, + "input_cost_per_token": 0.0, + "output_cost_per_token": 0.0, + "litellm_provider": "fireworks_ai", + "mode": "rerank" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-reranker-8b": { + "max_tokens": 40960, + "max_input_tokens": 40960, + "max_output_tokens": 40960, + "input_cost_per_token": 0.0, + "output_cost_per_token": 0.0, + "litellm_provider": "fireworks_ai", + "mode": "rerank" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-vl-235b-a22b-instruct": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 2.2e-07, + "output_cost_per_token": 8.8e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-vl-235b-a22b-thinking": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 2.2e-07, + "output_cost_per_token": 8.8e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-vl-30b-a3b-instruct": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 1.5e-07, + "output_cost_per_token": 6e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-vl-30b-a3b-thinking": { + "max_tokens": 262144, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "input_cost_per_token": 1.5e-07, + "output_cost_per_token": 6e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-vl-32b-instruct": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwen3-vl-8b-instruct": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/qwq-32b": { + "max_tokens": 131072, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/rolm-ocr": { + "max_tokens": 128000, + "max_input_tokens": 128000, + "max_output_tokens": 128000, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/snorkel-mistral-7b-pairrm-dpo": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/stable-diffusion-xl-1024-v1-0": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1.3e-10, + "output_cost_per_token": 1.3e-10, + "litellm_provider": "fireworks_ai", + "mode": "image_generation" + }, + "fireworks_ai/accounts/fireworks/models/stablecode-3b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/starcoder-16b": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/starcoder-7b": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/starcoder2-15b": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/starcoder2-3b": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/starcoder2-7b": { + "max_tokens": 16384, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/toppy-m-7b": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/whisper-v3": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 0.0, + "output_cost_per_token": 0.0, + "litellm_provider": "fireworks_ai", + "mode": "audio_transcription" + }, + "fireworks_ai/accounts/fireworks/models/whisper-v3-turbo": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 0.0, + "output_cost_per_token": 0.0, + "litellm_provider": "fireworks_ai", + "mode": "audio_transcription" + }, + "fireworks_ai/accounts/fireworks/models/yi-34b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/yi-34b-200k-capybara": { + "max_tokens": 200000, + "max_input_tokens": 200000, + "max_output_tokens": 200000, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/yi-34b-chat": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 9e-07, + "output_cost_per_token": 9e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/yi-6b": { + "max_tokens": 4096, + "max_input_tokens": 4096, + "max_output_tokens": 4096, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "fireworks_ai/accounts/fireworks/models/zephyr-7b-beta": { + "max_tokens": 32768, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "input_cost_per_token": 2e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "fireworks_ai", + "mode": "chat" + }, + "novita/deepseek/deepseek-v3.2": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2.69e-07, + "output_cost_per_token": 4e-07, + "max_input_tokens": 163840, + "max_output_tokens": 65536, + "max_tokens": 65536, "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "cache_read_input_token_cost": 1.345e-07, + "input_cost_per_token_cache_hit": 1.345e-07, + "supports_reasoning": true + }, + "novita/minimax/minimax-m2.1": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3e-07, + "output_cost_per_token": 1.2e-06, + "max_input_tokens": 204800, + "max_output_tokens": 131072, + "max_tokens": 131072, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "cache_read_input_token_cost": 3e-08, + "input_cost_per_token_cache_hit": 3e-08 + }, + "novita/zai-org/glm-4.7": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 6e-07, + "output_cost_per_token": 2.2e-06, + "max_input_tokens": 204800, + "max_output_tokens": 131072, + "max_tokens": 131072, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "cache_read_input_token_cost": 1.1e-07, + "input_cost_per_token_cache_hit": 1.1e-07, + "supports_reasoning": true + }, + "novita/xiaomimimo/mimo-v2-flash": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1e-07, + "output_cost_per_token": 3e-07, + "max_input_tokens": 262144, + "max_output_tokens": 32000, + "max_tokens": 32000, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "cache_read_input_token_cost": 2e-08, + "input_cost_per_token_cache_hit": 2e-08, + "supports_reasoning": true + }, + "novita/zai-org/autoglm-phone-9b-multilingual": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3.5e-08, + "output_cost_per_token": 1.38e-07, + "max_input_tokens": 65536, + "max_output_tokens": 65536, + "max_tokens": 65536, + "supports_vision": true, + "supports_system_messages": true + }, + "novita/moonshotai/kimi-k2-thinking": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 6e-07, + "output_cost_per_token": 2.5e-06, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "supports_reasoning": true + }, + "novita/minimax/minimax-m2": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3e-07, + "output_cost_per_token": 1.2e-06, + "max_input_tokens": 204800, + "max_output_tokens": 131072, + "max_tokens": 131072, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "cache_read_input_token_cost": 3e-08, + "input_cost_per_token_cache_hit": 3e-08, + "supports_reasoning": true + }, + "novita/paddlepaddle/paddleocr-vl": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2e-08, + "output_cost_per_token": 2e-08, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "max_tokens": 16384, + "supports_vision": true, + "supports_system_messages": true + }, + "novita/deepseek/deepseek-v3.2-exp": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2.7e-07, + "output_cost_per_token": 4.1e-07, + "max_input_tokens": 163840, + "max_output_tokens": 65536, + "max_tokens": 65536, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "supports_reasoning": true + }, + "novita/qwen/qwen3-vl-235b-a22b-thinking": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 9.8e-07, + "output_cost_per_token": 3.95e-06, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_vision": true, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/zai-org/glm-4.6v": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3e-07, + "output_cost_per_token": 9e-07, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, "supports_tool_choice": true, "supports_vision": true, - "source": "https://docs.z.ai/guides/overview/pricing" + "supports_system_messages": true, + "supports_response_schema": true, + "cache_read_input_token_cost": 5.5e-08, + "input_cost_per_token_cache_hit": 5.5e-08, + "supports_reasoning": true }, - "zai/glm-4.6": { - "cache_creation_input_token_cost": 0, + "novita/zai-org/glm-4.6": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 5.5e-07, + "output_cost_per_token": 2.2e-06, + "max_input_tokens": 204800, + "max_output_tokens": 131072, + "max_tokens": 131072, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, "cache_read_input_token_cost": 1.1e-07, + "input_cost_per_token_cache_hit": 1.1e-07, + "supports_reasoning": true + }, + "novita/kwaipilot/kat-coder-pro": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3e-07, + "output_cost_per_token": 1.2e-06, + "max_input_tokens": 256000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "cache_read_input_token_cost": 6e-08, + "input_cost_per_token_cache_hit": 6e-08 + }, + "novita/qwen/qwen3-next-80b-a3b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1.5e-07, + "output_cost_per_token": 1.5e-06, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/qwen/qwen3-next-80b-a3b-thinking": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1.5e-07, + "output_cost_per_token": 1.5e-06, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "supports_reasoning": true + }, + "novita/deepseek/deepseek-ocr": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3e-08, + "output_cost_per_token": 3e-08, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/deepseek/deepseek-v3.1-terminus": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2.7e-07, + "output_cost_per_token": 1e-06, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "cache_read_input_token_cost": 1.35e-07, + "input_cost_per_token_cache_hit": 1.35e-07, + "supports_reasoning": true + }, + "novita/qwen/qwen3-vl-235b-a22b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3e-07, + "output_cost_per_token": 1.5e-06, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/qwen/qwen3-max": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2.11e-06, + "output_cost_per_token": 8.45e-06, + "max_input_tokens": 262144, + "max_output_tokens": 65536, + "max_tokens": 65536, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/skywork/r1v4-lite": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2e-07, + "output_cost_per_token": 6e-07, + "max_input_tokens": 262144, + "max_output_tokens": 65536, + "max_tokens": 65536, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/deepseek/deepseek-v3.1": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2.7e-07, + "output_cost_per_token": 1e-06, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "cache_read_input_token_cost": 1.35e-07, + "input_cost_per_token_cache_hit": 1.35e-07, + "supports_reasoning": true + }, + "novita/moonshotai/kimi-k2-0905": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 6e-07, + "output_cost_per_token": 2.5e-06, + "max_input_tokens": 262144, + "max_output_tokens": 262144, + "max_tokens": 262144, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/qwen/qwen3-coder-480b-a35b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3e-07, + "output_cost_per_token": 1.3e-06, + "max_input_tokens": 262144, + "max_output_tokens": 65536, + "max_tokens": 65536, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/qwen/qwen3-coder-30b-a3b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 7e-08, + "output_cost_per_token": 2.7e-07, + "max_input_tokens": 160000, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/openai/gpt-oss-120b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 5e-08, + "output_cost_per_token": 2.5e-07, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true, + "supports_reasoning": true + }, + "novita/moonshotai/kimi-k2-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 5.7e-07, + "output_cost_per_token": 2.3e-06, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "max_tokens": 131072, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/deepseek/deepseek-v3-0324": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2.7e-07, + "output_cost_per_token": 1.12e-06, + "max_input_tokens": 163840, + "max_output_tokens": 163840, + "max_tokens": 163840, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "cache_read_input_token_cost": 1.35e-07, + "input_cost_per_token_cache_hit": 1.35e-07 + }, + "novita/zai-org/glm-4.5": { + "litellm_provider": "novita", + "mode": "chat", "input_cost_per_token": 6e-07, "output_cost_per_token": 2.2e-06, - "litellm_provider": "zai", - "max_input_tokens": 200000, - "max_output_tokens": 128000, - "mode": "chat", + "max_input_tokens": 131072, + "max_output_tokens": 98304, + "max_tokens": 98304, "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, + "supports_parallel_function_calling": true, "supports_tool_choice": true, - "source": "https://docs.z.ai/guides/overview/pricing" - }, - "zai/glm-4.7": { - "cache_creation_input_token_cost": 0, + "supports_system_messages": true, "cache_read_input_token_cost": 1.1e-07, + "input_cost_per_token_cache_hit": 1.1e-07, + "supports_reasoning": true + }, + "novita/qwen/qwen3-235b-a22b-thinking-2507": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3e-07, + "output_cost_per_token": 3e-06, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/meta-llama/llama-3.1-8b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2e-08, + "output_cost_per_token": 5e-08, + "max_input_tokens": 16384, + "max_output_tokens": 16384, + "max_tokens": 16384, + "supports_system_messages": true + }, + "novita/google/gemma-3-12b-it": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 5e-08, + "output_cost_per_token": 1e-07, + "max_input_tokens": 131072, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/zai-org/glm-4.5v": { + "litellm_provider": "novita", + "mode": "chat", "input_cost_per_token": 6e-07, + "output_cost_per_token": 1.8e-06, + "max_input_tokens": 65536, + "max_output_tokens": 16384, + "max_tokens": 16384, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true, + "cache_read_input_token_cost": 1.1e-07, + "input_cost_per_token_cache_hit": 1.1e-07, + "supports_reasoning": true + }, + "novita/openai/gpt-oss-20b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 4e-08, + "output_cost_per_token": 1.5e-07, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true, + "supports_reasoning": true + }, + "novita/qwen/qwen3-235b-a22b-instruct-2507": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 9e-08, + "output_cost_per_token": 5.8e-07, + "max_input_tokens": 131072, + "max_output_tokens": 16384, + "max_tokens": 16384, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/deepseek/deepseek-r1-distill-qwen-14b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1.5e-07, + "output_cost_per_token": 1.5e-07, + "max_input_tokens": 32768, + "max_output_tokens": 16384, + "max_tokens": 16384, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "supports_reasoning": true + }, + "novita/meta-llama/llama-3.3-70b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1.35e-07, + "output_cost_per_token": 4e-07, + "max_input_tokens": 131072, + "max_output_tokens": 120000, + "max_tokens": 120000, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true + }, + "novita/qwen/qwen-2.5-72b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3.8e-07, + "output_cost_per_token": 4e-07, + "max_input_tokens": 32000, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/mistralai/mistral-nemo": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 4e-08, + "output_cost_per_token": 1.7e-07, + "max_input_tokens": 60288, + "max_output_tokens": 16000, + "max_tokens": 16000, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/minimaxai/minimax-m1-80k": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 5.5e-07, "output_cost_per_token": 2.2e-06, - "litellm_provider": "zai", - "max_input_tokens": 200000, - "max_output_tokens": 128000, + "max_input_tokens": 1000000, + "max_output_tokens": 40000, + "max_tokens": 40000, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/deepseek/deepseek-r1-0528": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 7e-07, + "output_cost_per_token": 2.5e-06, + "max_input_tokens": 163840, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "cache_read_input_token_cost": 3.5e-07, + "input_cost_per_token_cache_hit": 3.5e-07, + "supports_reasoning": true + }, + "novita/deepseek/deepseek-r1-distill-qwen-32b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3e-07, + "output_cost_per_token": 3e-07, + "max_input_tokens": 64000, + "max_output_tokens": 32000, + "max_tokens": 32000, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "supports_reasoning": true + }, + "novita/meta-llama/llama-3-8b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 4e-08, + "output_cost_per_token": 4e-08, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_system_messages": true + }, + "novita/microsoft/wizardlm-2-8x22b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 6.2e-07, + "output_cost_per_token": 6.2e-07, + "max_input_tokens": 65535, + "max_output_tokens": 8000, + "max_tokens": 8000, + "supports_system_messages": true + }, + "novita/deepseek/deepseek-r1-0528-qwen3-8b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 6e-08, + "output_cost_per_token": 9e-08, + "max_input_tokens": 128000, + "max_output_tokens": 32000, + "max_tokens": 32000, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/deepseek/deepseek-r1-distill-llama-70b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 8e-07, + "output_cost_per_token": 8e-07, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true, + "supports_reasoning": true + }, + "novita/meta-llama/llama-3-70b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 5.1e-07, + "output_cost_per_token": 7.4e-07, + "max_input_tokens": 8192, + "max_output_tokens": 8000, + "max_tokens": 8000, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/qwen/qwen3-235b-a22b-fp8": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2e-07, + "output_cost_per_token": 8e-07, + "max_input_tokens": 40960, + "max_output_tokens": 20000, + "max_tokens": 20000, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/meta-llama/llama-4-maverick-17b-128e-instruct-fp8": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2.7e-07, + "output_cost_per_token": 8.5e-07, + "max_input_tokens": 1048576, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_vision": true, + "supports_system_messages": true + }, + "novita/meta-llama/llama-4-scout-17b-16e-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1.8e-07, + "output_cost_per_token": 5.9e-07, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "max_tokens": 131072, + "supports_vision": true, + "supports_system_messages": true + }, + "novita/nousresearch/hermes-2-pro-llama-3-8b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1.4e-07, + "output_cost_per_token": 1.4e-07, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/qwen/qwen2.5-vl-72b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 8e-07, + "output_cost_per_token": 8e-07, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_vision": true, + "supports_system_messages": true + }, + "novita/sao10k/l3-70b-euryale-v2.1": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1.48e-06, + "output_cost_per_token": 1.48e-06, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true + }, + "novita/baidu/ernie-4.5-21B-a3b-thinking": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 7e-08, + "output_cost_per_token": 2.8e-07, + "max_input_tokens": 131072, + "max_output_tokens": 65536, + "max_tokens": 65536, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/sao10k/l3-8b-lunaris": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 5e-08, + "output_cost_per_token": 5e-08, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/baichuan/baichuan-m2-32b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 7e-08, + "output_cost_per_token": 7e-08, + "max_input_tokens": 131072, + "max_output_tokens": 131072, + "max_tokens": 131072, + "supports_system_messages": true + }, + "novita/baidu/ernie-4.5-vl-424b-a47b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 4.2e-07, + "output_cost_per_token": 1.25e-06, + "max_input_tokens": 123000, + "max_output_tokens": 16000, + "max_tokens": 16000, + "supports_vision": true, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/baidu/ernie-4.5-300b-a47b-paddle": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2.8e-07, + "output_cost_per_token": 1.1e-06, + "max_input_tokens": 123000, + "max_output_tokens": 12000, + "max_tokens": 12000, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/deepseek/deepseek-prover-v2-671b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 7e-07, + "output_cost_per_token": 2.5e-06, + "max_input_tokens": 160000, + "max_output_tokens": 160000, + "max_tokens": 160000, + "supports_system_messages": true + }, + "novita/qwen/qwen3-32b-fp8": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1e-07, + "output_cost_per_token": 4.5e-07, + "max_input_tokens": 40960, + "max_output_tokens": 20000, + "max_tokens": 20000, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/qwen/qwen3-30b-a3b-fp8": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 9e-08, + "output_cost_per_token": 4.5e-07, + "max_input_tokens": 40960, + "max_output_tokens": 20000, + "max_tokens": 20000, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/google/gemma-3-27b-it": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1.19e-07, + "output_cost_per_token": 2e-07, + "max_input_tokens": 98304, + "max_output_tokens": 16384, + "max_tokens": 16384, + "supports_vision": true, + "supports_system_messages": true + }, + "novita/deepseek/deepseek-v3-turbo": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 4e-07, + "output_cost_per_token": 1.3e-06, + "max_input_tokens": 64000, + "max_output_tokens": 16000, + "max_tokens": 16000, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true + }, + "novita/deepseek/deepseek-r1-turbo": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 7e-07, + "output_cost_per_token": 2.5e-06, + "max_input_tokens": 64000, + "max_output_tokens": 16000, + "max_tokens": 16000, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/Sao10K/L3-8B-Stheno-v3.2": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 5e-08, + "output_cost_per_token": 5e-08, + "max_input_tokens": 8192, + "max_output_tokens": 32000, + "max_tokens": 32000, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true + }, + "novita/gryphe/mythomax-l2-13b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 9e-08, + "output_cost_per_token": 9e-08, + "max_input_tokens": 4096, + "max_output_tokens": 3200, + "max_tokens": 3200, + "supports_system_messages": true + }, + "novita/baidu/ernie-4.5-vl-28b-a3b-thinking": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3.9e-07, + "output_cost_per_token": 3.9e-07, + "max_input_tokens": 131072, + "max_output_tokens": 65536, + "max_tokens": 65536, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true, + "supports_reasoning": true + }, + "novita/qwen/qwen3-vl-8b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 8e-08, + "output_cost_per_token": 5e-07, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/zai-org/glm-4.5-air": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1.3e-07, + "output_cost_per_token": 8.5e-07, + "max_input_tokens": 131072, + "max_output_tokens": 98304, + "max_tokens": 98304, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/qwen/qwen3-vl-30b-a3b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2e-07, + "output_cost_per_token": 7e-07, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/qwen/qwen3-vl-30b-a3b-thinking": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2e-07, + "output_cost_per_token": 1e-06, + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/qwen/qwen3-omni-30b-a3b-thinking": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2.5e-07, + "output_cost_per_token": 9.7e-07, + "max_input_tokens": 65536, + "max_output_tokens": 16384, + "max_tokens": 16384, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true, + "supports_reasoning": true, + "supports_audio_input": true + }, + "novita/qwen/qwen3-omni-30b-a3b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2.5e-07, + "output_cost_per_token": 9.7e-07, + "max_input_tokens": 65536, + "max_output_tokens": 16384, + "max_tokens": 16384, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_response_schema": true, + "supports_audio_input": true, + "supports_audio_output": true + }, + "novita/qwen/qwen-mt-plus": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 2.5e-07, + "output_cost_per_token": 7.5e-07, + "max_input_tokens": 16384, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_system_messages": true + }, + "novita/baidu/ernie-4.5-vl-28b-a3b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1.4e-07, + "output_cost_per_token": 5.6e-07, + "max_input_tokens": 30000, + "max_output_tokens": 8000, + "max_tokens": 8000, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/baidu/ernie-4.5-21B-a3b": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 7e-08, + "output_cost_per_token": 2.8e-07, + "max_input_tokens": 120000, + "max_output_tokens": 8000, + "max_tokens": 8000, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true + }, + "novita/qwen/qwen3-8b-fp8": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3.5e-08, + "output_cost_per_token": 1.38e-07, + "max_input_tokens": 128000, + "max_output_tokens": 20000, + "max_tokens": 20000, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/qwen/qwen3-4b-fp8": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3e-08, + "output_cost_per_token": 3e-08, + "max_input_tokens": 128000, + "max_output_tokens": 20000, + "max_tokens": 20000, + "supports_system_messages": true, + "supports_reasoning": true + }, + "novita/qwen/qwen2.5-7b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 7e-08, + "output_cost_per_token": 7e-08, + "max_input_tokens": 32000, + "max_output_tokens": 32000, + "max_tokens": 32000, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true, + "supports_response_schema": true + }, + "novita/meta-llama/llama-3.2-3b-instruct": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 3e-08, + "output_cost_per_token": 5e-08, + "max_input_tokens": 32768, + "max_output_tokens": 32000, + "max_tokens": 32000, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true + }, + "novita/sao10k/l31-70b-euryale-v2.2": { + "litellm_provider": "novita", + "mode": "chat", + "input_cost_per_token": 1.48e-06, + "output_cost_per_token": 1.48e-06, + "max_input_tokens": 8192, + "max_output_tokens": 8192, + "max_tokens": 8192, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_tool_choice": true, + "supports_system_messages": true + }, + "novita/qwen/qwen3-embedding-0.6b": { + "litellm_provider": "novita", + "mode": "embedding", + "input_cost_per_token": 7e-08, + "output_cost_per_token": 0, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "max_tokens": 32768 + }, + "novita/qwen/qwen3-embedding-8b": { + "litellm_provider": "novita", + "mode": "embedding", + "input_cost_per_token": 7e-08, + "output_cost_per_token": 0, + "max_input_tokens": 32768, + "max_output_tokens": 4096, + "max_tokens": 4096 + }, + "novita/baai/bge-m3": { + "litellm_provider": "novita", + "mode": "embedding", + "input_cost_per_token": 1e-08, + "output_cost_per_token": 1e-08, + "max_input_tokens": 8192, + "max_output_tokens": 96000, + "max_tokens": 96000 + }, + "novita/qwen/qwen3-reranker-8b": { + "litellm_provider": "novita", + "mode": "rerank", + "input_cost_per_token": 5e-08, + "output_cost_per_token": 5e-08, + "max_input_tokens": 32768, + "max_output_tokens": 4096, + "max_tokens": 4096 + }, + "novita/baai/bge-reranker-v2-m3": { + "litellm_provider": "novita", + "mode": "rerank", + "input_cost_per_token": 1e-08, + "output_cost_per_token": 1e-08, + "max_input_tokens": 8000, + "max_output_tokens": 8000, + "max_tokens": 8000 + }, + "llamagate/llama-3.1-8b": { + "max_tokens": 8192, + "max_input_tokens": 131072, + "max_output_tokens": 8192, + "input_cost_per_token": 3e-08, + "output_cost_per_token": 5e-08, + "litellm_provider": "llamagate", "mode": "chat", "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_tool_choice": true, - "source": "https://docs.z.ai/guides/overview/pricing" + "supports_response_schema": true }, - "zai/glm-5": { - "cache_creation_input_token_cost": 0, - "cache_read_input_token_cost": 2e-07, - "input_cost_per_token": 1e-06, - "output_cost_per_token": 3.2e-06, - "litellm_provider": "zai", - "max_input_tokens": 200000, - "max_output_tokens": 128000, + "llamagate/llama-3.2-3b": { + "max_tokens": 8192, + "max_input_tokens": 131072, + "max_output_tokens": 8192, + "input_cost_per_token": 4e-08, + "output_cost_per_token": 8e-08, + "litellm_provider": "llamagate", "mode": "chat", "supports_function_calling": true, - "supports_prompt_caching": true, - "supports_reasoning": true, - "supports_tool_choice": true, - "source": "https://docs.z.ai/guides/overview/pricing" + "supports_response_schema": true }, - "zai/glm-5-code": { + "llamagate/mistral-7b-v0.3": { + "max_tokens": 8192, + "max_input_tokens": 32768, + "max_output_tokens": 8192, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 1.5e-07, + "litellm_provider": "llamagate", + "mode": "chat", + "supports_function_calling": true, + "supports_response_schema": true + }, + "llamagate/qwen3-8b": { + "max_tokens": 8192, + "max_input_tokens": 32768, + "max_output_tokens": 8192, + "input_cost_per_token": 4e-08, + "output_cost_per_token": 1.4e-07, + "litellm_provider": "llamagate", + "mode": "chat", + "supports_function_calling": true, + "supports_response_schema": true + }, + "llamagate/dolphin3-8b": { + "max_tokens": 8192, + "max_input_tokens": 128000, + "max_output_tokens": 8192, + "input_cost_per_token": 8e-08, + "output_cost_per_token": 1.5e-07, + "litellm_provider": "llamagate", + "mode": "chat", + "supports_function_calling": true, + "supports_response_schema": true + }, + "llamagate/deepseek-r1-8b": { + "max_tokens": 16384, + "max_input_tokens": 65536, + "max_output_tokens": 16384, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "llamagate", + "mode": "chat", + "supports_function_calling": true, + "supports_response_schema": true, + "supports_reasoning": true + }, + "llamagate/deepseek-r1-7b-qwen": { + "max_tokens": 16384, + "max_input_tokens": 131072, + "max_output_tokens": 16384, + "input_cost_per_token": 8e-08, + "output_cost_per_token": 1.5e-07, + "litellm_provider": "llamagate", + "mode": "chat", + "supports_function_calling": true, + "supports_response_schema": true, + "supports_reasoning": true + }, + "llamagate/openthinker-7b": { + "max_tokens": 8192, + "max_input_tokens": 32768, + "max_output_tokens": 8192, + "input_cost_per_token": 8e-08, + "output_cost_per_token": 1.5e-07, + "litellm_provider": "llamagate", + "mode": "chat", + "supports_function_calling": true, + "supports_response_schema": true, + "supports_reasoning": true + }, + "llamagate/qwen2.5-coder-7b": { + "max_tokens": 8192, + "max_input_tokens": 32768, + "max_output_tokens": 8192, + "input_cost_per_token": 6e-08, + "output_cost_per_token": 1.2e-07, + "litellm_provider": "llamagate", + "mode": "chat", + "supports_function_calling": true, + "supports_response_schema": true + }, + "llamagate/deepseek-coder-6.7b": { + "max_tokens": 4096, + "max_input_tokens": 16384, + "max_output_tokens": 4096, + "input_cost_per_token": 6e-08, + "output_cost_per_token": 1.2e-07, + "litellm_provider": "llamagate", + "mode": "chat", + "supports_function_calling": true, + "supports_response_schema": true + }, + "llamagate/codellama-7b": { + "max_tokens": 4096, + "max_input_tokens": 16384, + "max_output_tokens": 4096, + "input_cost_per_token": 6e-08, + "output_cost_per_token": 1.2e-07, + "litellm_provider": "llamagate", + "mode": "chat", + "supports_function_calling": true, + "supports_response_schema": true + }, + "llamagate/qwen3-vl-8b": { + "max_tokens": 8192, + "max_input_tokens": 32768, + "max_output_tokens": 8192, + "input_cost_per_token": 1.5e-07, + "output_cost_per_token": 5.5e-07, + "litellm_provider": "llamagate", + "mode": "chat", + "supports_function_calling": true, + "supports_response_schema": true, + "supports_vision": true + }, + "llamagate/llava-7b": { + "max_tokens": 2048, + "max_input_tokens": 4096, + "max_output_tokens": 2048, + "input_cost_per_token": 1e-07, + "output_cost_per_token": 2e-07, + "litellm_provider": "llamagate", + "mode": "chat", + "supports_response_schema": true, + "supports_vision": true + }, + "llamagate/gemma3-4b": { + "max_tokens": 8192, + "max_input_tokens": 128000, + "max_output_tokens": 8192, + "input_cost_per_token": 3e-08, + "output_cost_per_token": 8e-08, + "litellm_provider": "llamagate", + "mode": "chat", + "supports_function_calling": true, + "supports_response_schema": true, + "supports_vision": true + }, + "llamagate/nomic-embed-text": { + "max_tokens": 8192, + "max_input_tokens": 8192, + "input_cost_per_token": 2e-08, + "output_cost_per_token": 0, + "litellm_provider": "llamagate", + "mode": "embedding" + }, + "llamagate/qwen3-embedding-8b": { + "max_tokens": 40960, + "max_input_tokens": 40960, + "input_cost_per_token": 2e-08, + "output_cost_per_token": 0, + "litellm_provider": "llamagate", + "mode": "embedding" + }, + "sarvam/sarvam-m": { "cache_creation_input_token_cost": 0, - "cache_read_input_token_cost": 3e-07, - "input_cost_per_token": 1.2e-06, + "cache_creation_input_token_cost_above_1hr": 0, + "cache_read_input_token_cost": 0, + "input_cost_per_token": 0, + "litellm_provider": "sarvam", + "max_input_tokens": 8192, + "max_output_tokens": 32000, + "max_tokens": 32000, + "mode": "chat", + "output_cost_per_token": 0, + "supports_reasoning": true + }, + "tts-1-1106": { + "input_cost_per_character": 1.5e-05, + "litellm_provider": "openai", + "mode": "audio_speech", + "supported_endpoints": [ + "/v1/audio/speech" + ] + }, + "tts-1-hd-1106": { + "input_cost_per_character": 3e-05, + "litellm_provider": "openai", + "mode": "audio_speech", + "supported_endpoints": [ + "/v1/audio/speech" + ] + }, + "gpt-4o-mini-tts-2025-03-20": { + "input_cost_per_token": 2.5e-06, + "litellm_provider": "openai", + "mode": "audio_speech", + "output_cost_per_audio_token": 1.2e-05, + "output_cost_per_second": 0.00025, + "output_cost_per_token": 1e-05, + "supported_endpoints": [ + "/v1/audio/speech" + ], + "supported_modalities": [ + "text", + "audio" + ], + "supported_output_modalities": [ + "audio" + ] + }, + "gpt-4o-mini-tts-2025-12-15": { + "input_cost_per_token": 2.5e-06, + "litellm_provider": "openai", + "mode": "audio_speech", + "output_cost_per_audio_token": 1.2e-05, + "output_cost_per_second": 0.00025, + "output_cost_per_token": 1e-05, + "supported_endpoints": [ + "/v1/audio/speech" + ], + "supported_modalities": [ + "text", + "audio" + ], + "supported_output_modalities": [ + "audio" + ] + }, + "gpt-4o-mini-transcribe-2025-03-20": { + "input_cost_per_audio_token": 3e-06, + "input_cost_per_token": 1.25e-06, + "litellm_provider": "openai", + "max_input_tokens": 16000, + "max_output_tokens": 2000, + "mode": "audio_transcription", "output_cost_per_token": 5e-06, - "litellm_provider": "zai", - "max_input_tokens": 200000, + "supported_endpoints": [ + "/v1/audio/transcriptions" + ] + }, + "gpt-4o-mini-transcribe-2025-12-15": { + "input_cost_per_audio_token": 3e-06, + "input_cost_per_token": 1.25e-06, + "litellm_provider": "openai", + "max_input_tokens": 16000, + "max_output_tokens": 2000, + "mode": "audio_transcription", + "output_cost_per_token": 5e-06, + "supported_endpoints": [ + "/v1/audio/transcriptions" + ] + }, + "gpt-5-search-api": { + "cache_read_input_token_cost": 1.25e-07, + "input_cost_per_token": 1.25e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, "max_output_tokens": 128000, + "max_tokens": 128000, "mode": "chat", + "output_cost_per_token": 1e-05, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-5-search-api-2025-10-14": { + "cache_read_input_token_cost": 1.25e-07, + "input_cost_per_token": 1.25e-06, + "litellm_provider": "openai", + "max_input_tokens": 272000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1e-05, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true, + "supports_web_search": true, + "supports_none_reasoning_effort": false, + "supports_xhigh_reasoning_effort": false + }, + "gpt-realtime-mini-2025-10-06": { + "cache_creation_input_audio_token_cost": 3e-07, + "cache_read_input_audio_token_cost": 3e-07, + "cache_read_input_token_cost": 6e-08, + "input_cost_per_audio_token": 1e-05, + "input_cost_per_image": 8e-07, + "input_cost_per_token": 6e-07, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_audio_token": 2e-05, + "output_cost_per_token": 2.4e-06, + "supported_endpoints": [ + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "image", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "gpt-realtime-mini-2025-12-15": { + "cache_creation_input_audio_token_cost": 3e-07, + "cache_read_input_audio_token_cost": 3e-07, + "cache_read_input_token_cost": 6e-08, + "input_cost_per_audio_token": 1e-05, + "input_cost_per_image": 8e-07, + "input_cost_per_token": 6e-07, + "litellm_provider": "openai", + "max_input_tokens": 128000, + "max_output_tokens": 4096, + "max_tokens": 4096, + "mode": "chat", + "output_cost_per_audio_token": 2e-05, + "output_cost_per_token": 2.4e-06, + "supported_endpoints": [ + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "image", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_system_messages": true, + "supports_tool_choice": true + }, + "sora-2": { + "litellm_provider": "openai", + "mode": "video_generation", + "output_cost_per_video_per_second": 0.1, + "source": "https://platform.openai.com/docs/api-reference/videos", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "video" + ], + "supported_resolutions": [ + "720x1280", + "1280x720" + ] + }, + "sora-2-pro": { + "litellm_provider": "openai", + "mode": "video_generation", + "output_cost_per_video_per_second": 0.3, + "source": "https://platform.openai.com/docs/api-reference/videos", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "video" + ], + "supported_resolutions": [ + "720x1280", + "1280x720" + ] + }, + "sora-2-pro-high-res": { + "litellm_provider": "openai", + "mode": "video_generation", + "output_cost_per_video_per_second": 0.5, + "source": "https://platform.openai.com/docs/api-reference/videos", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "video" + ], + "supported_resolutions": [ + "1024x1792", + "1792x1024" + ] + }, + "chatgpt-image-latest": { + "cache_read_input_image_token_cost": 2.5e-06, + "cache_read_input_token_cost": 1.25e-06, + "input_cost_per_image_token": 1e-05, + "input_cost_per_token": 5e-06, + "litellm_provider": "openai", + "mode": "image_generation", + "output_cost_per_image_token": 4e-05, + "supported_endpoints": [ + "/v1/images/generations", + "/v1/images/edits" + ] + }, + "gemini-2.0-flash-exp-image-generation": { + "input_cost_per_token": 0.0, + "litellm_provider": "gemini", + "max_images_per_prompt": 3000, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "image_generation", + "output_cost_per_image": 0.039, + "output_cost_per_token": 0.0, + "source": "https://ai.google.dev/pricing", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_vision": true + }, + "gemini/gemini-2.0-flash-exp-image-generation": { + "input_cost_per_token": 0.0, + "litellm_provider": "gemini", + "max_images_per_prompt": 3000, + "max_input_tokens": 32768, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "image_generation", + "output_cost_per_image": 0.039, + "output_cost_per_token": 0.0, + "source": "https://ai.google.dev/pricing", + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_vision": true, + "tpm": 250000, + "rpm": 10 + }, + "gemini/gemini-2.0-flash-lite-001": { + "cache_read_input_token_cost": 1.875e-08, + "deprecation_date": "2026-06-01", + "input_cost_per_audio_token": 7.5e-08, + "input_cost_per_token": 7.5e-08, + "litellm_provider": "gemini", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 8192, + "max_pdf_size_mb": 50, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_token": 3e-07, + "rpm": 4000, + "source": "https://ai.google.dev/gemini-api/docs/pricing#gemini-2.0-flash-lite", + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_output": true, "supports_function_calling": true, "supports_prompt_caching": true, - "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, "supports_tool_choice": true, - "source": "https://docs.z.ai/guides/overview/pricing" + "supports_vision": true, + "supports_web_search": true, + "tpm": 4000000 + }, + "gemini-2.5-flash-native-audio-latest": { + "input_cost_per_audio_token": 1e-06, + "input_cost_per_token": 3e-07, + "litellm_provider": "gemini", + "max_input_tokens": 1048576, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 2.5e-06, + "source": "https://ai.google.dev/pricing", + "supported_endpoints": [ + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true + }, + "gemini-2.5-flash-native-audio-preview-09-2025": { + "input_cost_per_audio_token": 1e-06, + "input_cost_per_token": 3e-07, + "litellm_provider": "gemini", + "max_input_tokens": 1048576, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 2.5e-06, + "source": "https://ai.google.dev/pricing", + "supported_endpoints": [ + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true + }, + "gemini-2.5-flash-native-audio-preview-12-2025": { + "input_cost_per_audio_token": 1e-06, + "input_cost_per_token": 3e-07, + "litellm_provider": "gemini", + "max_input_tokens": 1048576, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 2.5e-06, + "source": "https://ai.google.dev/pricing", + "supported_endpoints": [ + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true + }, + "gemini/gemini-2.5-flash-native-audio-latest": { + "input_cost_per_audio_token": 1e-06, + "input_cost_per_token": 3e-07, + "litellm_provider": "gemini", + "max_input_tokens": 1048576, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 2.5e-06, + "source": "https://ai.google.dev/pricing", + "supported_endpoints": [ + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "tpm": 250000, + "rpm": 10 + }, + "gemini/gemini-2.5-flash-native-audio-preview-09-2025": { + "input_cost_per_audio_token": 1e-06, + "input_cost_per_token": 3e-07, + "litellm_provider": "gemini", + "max_input_tokens": 1048576, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 2.5e-06, + "source": "https://ai.google.dev/pricing", + "supported_endpoints": [ + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "tpm": 250000, + "rpm": 10 + }, + "gemini/gemini-2.5-flash-native-audio-preview-12-2025": { + "input_cost_per_audio_token": 1e-06, + "input_cost_per_token": 3e-07, + "litellm_provider": "gemini", + "max_input_tokens": 1048576, + "max_output_tokens": 8192, + "max_tokens": 8192, + "mode": "chat", + "output_cost_per_token": 2.5e-06, + "source": "https://ai.google.dev/pricing", + "supported_endpoints": [ + "/v1/realtime" + ], + "supported_modalities": [ + "text", + "audio" + ], + "supported_output_modalities": [ + "text", + "audio" + ], + "supports_audio_input": true, + "supports_audio_output": true, + "tpm": 250000, + "rpm": 10 + }, + "gemini-2.5-flash-preview-tts": { + "input_cost_per_token": 3e-07, + "litellm_provider": "gemini", + "mode": "audio_speech", + "output_cost_per_token": 2.5e-06, + "source": "https://ai.google.dev/pricing", + "supported_endpoints": [ + "/v1/audio/speech" + ] + }, + "gemini-flash-latest": { + "cache_read_input_token_cost": 3e-08, + "input_cost_per_audio_token": 1e-06, + "input_cost_per_token": 3e-07, + "litellm_provider": "gemini", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_reasoning_token": 2.5e-06, + "output_cost_per_token": 2.5e-06, + "rpm": 100000, + "source": "https://ai.google.dev/gemini-api/docs/models#gemini-2.5-flash-preview", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true, + "tpm": 8000000 + }, + "gemini-flash-lite-latest": { + "cache_read_input_token_cost": 1e-08, + "input_cost_per_audio_token": 3e-07, + "input_cost_per_token": 1e-07, + "litellm_provider": "gemini", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_reasoning_token": 4e-07, + "output_cost_per_token": 4e-07, + "rpm": 15, + "source": "https://ai.google.dev/gemini-api/docs/models#gemini-2.5-flash-lite", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true, + "tpm": 250000 + }, + "gemini-pro-latest": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_above_200k_tokens": 2.5e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_above_200k_tokens": 2.5e-06, + "litellm_provider": "gemini", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_above_200k_tokens": 1.5e-05, + "rpm": 2000, + "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_input": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_video_input": true, + "supports_vision": true, + "supports_web_search": true, + "tpm": 800000 + }, + "gemini/gemini-pro-latest": { + "cache_read_input_token_cost": 1.25e-07, + "cache_read_input_token_cost_above_200k_tokens": 2.5e-07, + "input_cost_per_token": 1.25e-06, + "input_cost_per_token_above_200k_tokens": 2.5e-06, + "litellm_provider": "gemini", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_token": 1e-05, + "output_cost_per_token_above_200k_tokens": 1.5e-05, + "rpm": 2000, + "source": "https://cloud.google.com/vertex-ai/generative-ai/pricing", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_input": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_video_input": true, + "supports_vision": true, + "supports_web_search": true, + "tpm": 800000 + }, + "gemini-exp-1206": { + "cache_read_input_token_cost": 3e-08, + "input_cost_per_audio_token": 1e-06, + "input_cost_per_token": 3e-07, + "litellm_provider": "gemini", + "max_audio_length_hours": 8.4, + "max_audio_per_prompt": 1, + "max_images_per_prompt": 3000, + "max_input_tokens": 1048576, + "max_output_tokens": 65535, + "max_pdf_size_mb": 30, + "max_tokens": 65535, + "max_video_length": 1, + "max_videos_per_prompt": 10, + "mode": "chat", + "output_cost_per_reasoning_token": 2.5e-06, + "output_cost_per_token": 2.5e-06, + "rpm": 100000, + "source": "https://ai.google.dev/gemini-api/docs/models#gemini-2.5-flash-preview", + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/completions", + "/v1/batch" + ], + "supported_modalities": [ + "text", + "image", + "audio", + "video" + ], + "supported_output_modalities": [ + "text" + ], + "supports_audio_output": false, + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_url_context": true, + "supports_vision": true, + "supports_web_search": true, + "tpm": 8000000 + }, + "vertex_ai/claude-sonnet-4-6@default": { + "cache_creation_input_token_cost": 3.75e-06, + "cache_creation_input_token_cost_above_200k_tokens": 7.5e-06, + "cache_read_input_token_cost": 3e-07, + "cache_read_input_token_cost_above_200k_tokens": 6e-07, + "input_cost_per_token": 3e-06, + "input_cost_per_token_above_200k_tokens": 6e-06, + "litellm_provider": "vertex_ai-anthropic_models", + "max_input_tokens": 200000, + "max_output_tokens": 64000, + "max_tokens": 64000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "output_cost_per_token_above_200k_tokens": 2.25e-05, + "supports_assistant_prefill": true, + "supports_computer_use": true, + "supports_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true, + "supports_vision": true, + "tool_use_system_prompt_tokens": 346, + "search_context_cost_per_query": { + "search_context_size_high": 0.01, + "search_context_size_low": 0.01, + "search_context_size_medium": 0.01 + } + }, + "duckduckgo/search": { + "litellm_provider": "duckduckgo", + "mode": "search", + "input_cost_per_query": 0.0, + "metadata": { + "notes": "DuckDuckGo Instant Answer API is free and does not require an API key." + } + }, + "bedrock_mantle/openai.gpt-oss-120b": { + "input_cost_per_token": 1.5e-07, + "output_cost_per_token": 6e-07, + "litellm_provider": "bedrock_mantle", + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "chat", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "bedrock_mantle/openai.gpt-oss-20b": { + "input_cost_per_token": 7.5e-08, + "output_cost_per_token": 3e-07, + "litellm_provider": "bedrock_mantle", + "max_input_tokens": 131072, + "max_output_tokens": 32768, + "max_tokens": 32768, + "mode": "chat", + "supports_function_calling": true, + "supports_parallel_function_calling": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "bedrock_mantle/openai.gpt-oss-safeguard-120b": { + "input_cost_per_token": 1.5e-07, + "output_cost_per_token": 6e-07, + "litellm_provider": "bedrock_mantle", + "max_input_tokens": 131072, + "max_output_tokens": 65536, + "max_tokens": 65536, + "mode": "chat", + "supports_function_calling": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true + }, + "bedrock_mantle/openai.gpt-oss-safeguard-20b": { + "input_cost_per_token": 7.5e-08, + "output_cost_per_token": 3e-07, + "litellm_provider": "bedrock_mantle", + "max_input_tokens": 131072, + "max_output_tokens": 65536, + "max_tokens": 65536, + "mode": "chat", + "supports_function_calling": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_tool_choice": true } } diff --git a/tests/litellm/llms/vertex_ai/test_gemini_batch_embeddings.py b/tests/litellm/llms/vertex_ai/test_gemini_batch_embeddings.py index 7047be4241..1ed1de01b5 100644 --- a/tests/litellm/llms/vertex_ai/test_gemini_batch_embeddings.py +++ b/tests/litellm/llms/vertex_ai/test_gemini_batch_embeddings.py @@ -15,8 +15,16 @@ from unittest.mock import MagicMock, patch sys.path.insert(0, os.path.abspath("../../../..")) import pytest + import litellm from litellm.llms.custom_httpx.http_handler import HTTPHandler +from litellm.llms.vertex_ai.gemini_embeddings.batch_embed_content_transformation import ( + _is_multimodal_input, + _parse_data_url, + process_embed_content_response, + transform_openai_input_gemini_embed_content, +) +from litellm.types.utils import EmbeddingResponse def test_gemini_batch_embeddings_with_custom_api_base_and_auth_header(): @@ -47,11 +55,9 @@ def test_gemini_batch_embeddings_with_custom_api_base_and_auth_header(): mock_response = MagicMock() mock_response.status_code = 200 mock_response.json.return_value = { - "predictions": [ + "embeddings": [ { - "embeddings": { - "values": [0.1, 0.2, 0.3, 0.4, 0.5] - } + "values": [0.1, 0.2, 0.3, 0.4, 0.5] } ] } @@ -109,11 +115,9 @@ def test_gemini_batch_embeddings_with_extra_headers(): mock_response = MagicMock() mock_response.status_code = 200 mock_response.json.return_value = { - "predictions": [ + "embeddings": [ { - "embeddings": { - "values": [0.1, 0.2, 0.3] - } + "values": [0.1, 0.2, 0.3] } ] } @@ -143,3 +147,380 @@ def test_gemini_batch_embeddings_with_extra_headers(): assert "X-Custom" in headers assert headers["X-Custom"] == "custom-value" + +def test_is_multimodal_input_detection(): + """Test that _is_multimodal_input correctly detects multimodal inputs.""" + assert _is_multimodal_input("plain text") is False + assert _is_multimodal_input(["text1", "text2"]) is False + + assert _is_multimodal_input("data:image/png;base64,iVBORw0KGgo=") is True + assert _is_multimodal_input(["text", "data:image/png;base64,abc"]) is True + + assert _is_multimodal_input("files/abc123") is True + assert _is_multimodal_input(["text", "files/myfile"]) is True + + +def test_parse_data_url(): + """Test that _parse_data_url correctly extracts MIME type and base64 data.""" + mime_type, base64_data = _parse_data_url("data:image/png;base64,iVBORw0KGgo=") + assert mime_type == "image/png" + assert base64_data == "iVBORw0KGgo=" + + mime_type, base64_data = _parse_data_url("data:audio/mpeg;base64,SUQzBAA=") + assert mime_type == "audio/mpeg" + assert base64_data == "SUQzBAA=" + + mime_type, base64_data = _parse_data_url("data:video/mp4;base64,AAAAIGZ0eXA=") + assert mime_type == "video/mp4" + assert base64_data == "AAAAIGZ0eXA=" + + mime_type, base64_data = _parse_data_url("data:application/pdf;base64,JVBERi0=") + assert mime_type == "application/pdf" + assert base64_data == "JVBERi0=" + + +def test_mime_type_validation(): + """Test that unsupported MIME types raise ValueError.""" + with pytest.raises(ValueError, match="Unsupported MIME type"): + _parse_data_url("data:text/plain;base64,SGVsbG8=") + + with pytest.raises(ValueError, match="Unsupported MIME type"): + _parse_data_url("data:application/json;base64,e30=") + + +def test_parse_data_url_invalid_format(): + """Test that invalid data URL formats raise ValueError.""" + with pytest.raises(ValueError, match="Invalid data URL format"): + _parse_data_url("not-a-data-url") + + with pytest.raises(ValueError, match="missing comma"): + _parse_data_url("data:image/png;base64") + + +def test_transform_multimodal_text_and_image(): + """Test transformation of mixed text and image input.""" + input_data = [ + "The food was delicious", + "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg==" + ] + + result = transform_openai_input_gemini_embed_content( + input=input_data, + model="gemini-embedding-2-preview", + optional_params={}, + resolved_files=None, + ) + + assert "content" in result + assert "parts" in result["content"] + parts = result["content"]["parts"] + + assert len(parts) == 2 + assert parts[0]["text"] == "The food was delicious" + assert "inline_data" in parts[1] + assert parts[1]["inline_data"]["mime_type"] == "image/png" + assert "data" in parts[1]["inline_data"] + + +def test_transform_multimodal_with_file_reference(): + """Test transformation with Gemini file reference.""" + input_data = ["Some text", "files/abc123"] + + resolved_files = { + "files/abc123": { + "mime_type": "image/jpeg", + "uri": "https://generativelanguage.googleapis.com/v1beta/files/abc123" + } + } + + result = transform_openai_input_gemini_embed_content( + input=input_data, + model="gemini-embedding-2-preview", + optional_params={}, + resolved_files=resolved_files, + ) + + assert "content" in result + parts = result["content"]["parts"] + + assert len(parts) == 2 + assert parts[0]["text"] == "Some text" + assert "file_data" in parts[1] + assert parts[1]["file_data"]["mime_type"] == "image/jpeg" + assert parts[1]["file_data"]["file_uri"] == "https://generativelanguage.googleapis.com/v1beta/files/abc123" + + +def test_embed_content_response_processing(): + """Test processing of embedContent response (single embedding).""" + response_json = { + "embedding": { + "values": [0.1, 0.2, 0.3, 0.4, 0.5] + } + } + + model_response = EmbeddingResponse() + result = process_embed_content_response( + input=["test input"], + model_response=model_response, + model="gemini-embedding-2-preview", + response_json=response_json, + ) + + assert len(result.data) == 1 + assert result.data[0].embedding == [0.1, 0.2, 0.3, 0.4, 0.5] + assert result.data[0].index == 0 + assert result.data[0].object == "embedding" + assert result.model == "gemini-embedding-2-preview" + assert result.usage.prompt_tokens > 0 + + +def test_embed_content_response_multimodal_sets_prompt_tokens_zero(): + """Test that multimodal input sets prompt_tokens=0 (cannot accurately count).""" + response_json = { + "embedding": { + "values": [0.1, 0.2, 0.3, 0.4, 0.5] + } + } + + model_response = EmbeddingResponse() + result = process_embed_content_response( + input=["text", "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg=="], + model_response=model_response, + model="gemini-embedding-2-preview", + response_json=response_json, + ) + + assert result.usage.prompt_tokens == 0 + + +def test_gemini_multimodal_embedding_e2e(): + """Test end-to-end multimodal embedding call through litellm.embedding().""" + client = HTTPHandler() + + def mock_auth_token(*args, **kwargs): + return None, "test-project" + + with patch.object(client, "post") as mock_post, patch( + "litellm.llms.vertex_ai.gemini_embeddings.batch_embed_content_handler.GoogleBatchEmbeddings._ensure_access_token", + side_effect=mock_auth_token + ), patch( + "litellm.llms.vertex_ai.gemini_embeddings.batch_embed_content_handler.GoogleBatchEmbeddings._get_token_and_url" + ) as mock_get_token: + mock_get_token.return_value = ( + {"x-goog-api-key": "test-key"}, + "https://generativelanguage.googleapis.com/v1beta/models/gemini-embedding-2-preview:embedContent?key=test-key" + ) + + mock_response = MagicMock() + mock_response.status_code = 200 + mock_response.json.return_value = { + "embedding": { + "values": [0.1, 0.2, 0.3, 0.4, 0.5] + } + } + mock_post.return_value = mock_response + + response = litellm.embedding( + model="gemini/gemini-embedding-2-preview", + input=["The food was delicious", "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg=="], + api_key="test-key", + client=client + ) + + mock_post.assert_called_once() + + call_args = mock_post.call_args + kwargs = call_args.kwargs if hasattr(call_args, 'kwargs') else call_args[1] + + request_body = json.loads(kwargs.get("data", "{}")) + + assert "content" in request_body + assert "parts" in request_body["content"] + parts = request_body["content"]["parts"] + + assert len(parts) == 2 + assert parts[0]["text"] == "The food was delicious" + assert "inline_data" in parts[1] + assert parts[1]["inline_data"]["mime_type"] == "image/png" + + assert len(response.data) == 1 + assert response.data[0].embedding == [0.1, 0.2, 0.3, 0.4, 0.5] + + +def test_gemini_multimodal_embedding_with_audio(): + """Test multimodal embedding with audio input.""" + input_data = ["Audio description", "data:audio/mpeg;base64,SUQzBAAAAAA="] + + result = transform_openai_input_gemini_embed_content( + input=input_data, + model="gemini-embedding-2-preview", + optional_params={}, + resolved_files=None, + ) + + parts = result["content"]["parts"] + assert len(parts) == 2 + assert parts[0]["text"] == "Audio description" + assert parts[1]["inline_data"]["mime_type"] == "audio/mpeg" + + +def test_gemini_multimodal_embedding_with_video(): + """Test multimodal embedding with video input.""" + input_data = ["data:video/mp4;base64,AAAAIGZ0eXBpc29tAAACAGlzb21pc28yYXZjMW1wNDEAAAAIZnJlZQAA"] + + result = transform_openai_input_gemini_embed_content( + input=input_data, + model="gemini-embedding-2-preview", + optional_params={}, + resolved_files=None, + ) + + parts = result["content"]["parts"] + assert len(parts) == 1 + assert parts[0]["inline_data"]["mime_type"] == "video/mp4" + + + +def test_transform_with_optional_params(): + """Test that optional params like outputDimensionality are passed through.""" + input_data = ["test text"] + + result = transform_openai_input_gemini_embed_content( + input=input_data, + model="gemini-embedding-2-preview", + optional_params={"outputDimensionality": 768, "taskType": "SEMANTIC_SIMILARITY"}, + resolved_files=None, + ) + + assert result["outputDimensionality"] == 768 + assert result["taskType"] == "SEMANTIC_SIMILARITY" + + +def test_dimensions_mapped_to_output_dimensionality(): + """Test that OpenAI 'dimensions' param is mapped to Gemini 'outputDimensionality'.""" + input_data = ["test text"] + + result = transform_openai_input_gemini_embed_content( + input=input_data, + model="gemini-embedding-2-preview", + optional_params={"dimensions": 768}, + resolved_files=None, + ) + + assert "outputDimensionality" in result + assert result["outputDimensionality"] == 768 + assert "dimensions" not in result + + +def test_is_gcs_url(): + """Test GCS URL detection.""" + from litellm.llms.vertex_ai.gemini_embeddings.batch_embed_content_transformation import ( + _is_gcs_url, + ) + + assert _is_gcs_url("gs://my-bucket/path/to/file.png") is True + assert _is_gcs_url("gs://bucket/image.jpg") is True + assert _is_gcs_url("https://storage.googleapis.com/bucket/file.png") is False + assert _is_gcs_url("files/abc123") is False + assert _is_gcs_url("data:image/png;base64,abc") is False + assert _is_gcs_url("regular text") is False + + +def test_infer_mime_type_from_gcs_url(): + """Test MIME type inference from GCS URL.""" + from litellm.llms.vertex_ai.gemini_embeddings.batch_embed_content_transformation import ( + _infer_mime_type_from_gcs_url, + ) + + assert _infer_mime_type_from_gcs_url("gs://bucket/image.png") == "image/png" + assert _infer_mime_type_from_gcs_url("gs://bucket/photo.jpg") == "image/jpeg" + assert _infer_mime_type_from_gcs_url("gs://bucket/photo.JPEG") == "image/jpeg" + assert _infer_mime_type_from_gcs_url("gs://bucket/audio.mp3") == "audio/mpeg" + assert _infer_mime_type_from_gcs_url("gs://bucket/audio.wav") == "audio/wav" + assert _infer_mime_type_from_gcs_url("gs://bucket/video.mp4") == "video/mp4" + assert _infer_mime_type_from_gcs_url("gs://bucket/video.mov") == "video/quicktime" + assert _infer_mime_type_from_gcs_url("gs://bucket/doc.pdf") == "application/pdf" + + with pytest.raises(ValueError, match="Unable to infer MIME type"): + _infer_mime_type_from_gcs_url("gs://bucket/file.txt") + + +def test_transform_multimodal_with_gcs_url(): + """Test transformation with GCS URL.""" + input_data = [ + "Describe this image", + "gs://my-bucket/images/photo.png" + ] + + result = transform_openai_input_gemini_embed_content( + input=input_data, + model="gemini-embedding-2-preview", + optional_params={}, + resolved_files=None, + ) + + parts = result["content"]["parts"] + assert len(parts) == 2 + assert parts[0]["text"] == "Describe this image" + assert parts[1]["file_data"]["mime_type"] == "image/png" + assert parts[1]["file_data"]["file_uri"] == "gs://my-bucket/images/photo.png" + + +def test_multimodal_input_detection_with_gcs(): + """Test that GCS URLs are detected as multimodal.""" + from litellm.llms.vertex_ai.gemini_embeddings.batch_embed_content_transformation import ( + _is_multimodal_input, + ) + + assert _is_multimodal_input(["text", "gs://bucket/file.png"]) is True + assert _is_multimodal_input("gs://bucket/video.mp4") is True + assert _is_multimodal_input(["just text", "more text"]) is False + + +def test_vertex_ai_text_only_embedding_uses_embed_content(): + """ + Test that vertex_ai/gemini-embedding-2-preview with text-only input uses + embedContent endpoint (not batchEmbedContents) and returns a single embedding. + """ + client = HTTPHandler() + embed_content_url = "https://us-central1-aiplatform.googleapis.com/v1/projects/test/locations/us-central1/publishers/google/models/gemini-embedding-2-preview:embedContent" + + def mock_auth_token(*args, **kwargs): + return "Bearer test-token", "test-project" + + with patch.object(client, "post") as mock_post, patch( + "litellm.llms.vertex_ai.gemini_embeddings.batch_embed_content_handler.GoogleBatchEmbeddings._ensure_access_token", + side_effect=mock_auth_token, + ), patch( + "litellm.llms.vertex_ai.gemini_embeddings.batch_embed_content_handler.GoogleBatchEmbeddings._get_token_and_url" + ) as mock_get_token: + mock_get_token.return_value = ( + {"Authorization": "Bearer test-token"}, + embed_content_url, + ) + mock_response = MagicMock() + mock_response.status_code = 200 + mock_response.json.return_value = { + "embedding": {"values": [0.1, 0.2, 0.3, 0.4, 0.5]} + } + mock_post.return_value = mock_response + + response = litellm.embedding( + model="vertex_ai/gemini-embedding-2-preview", + input=["Hello, world!"], + vertex_project="test-project", + vertex_location="us-central1", + client=client, + ) + + mock_post.assert_called_once() + call_args = mock_post.call_args + post_url = call_args.kwargs.get("url", call_args.args[0] if call_args.args else "") + assert "embedContent" in str(post_url) + data = json.loads(call_args.kwargs["data"]) + assert "content" in data + assert "parts" in data["content"] + assert len(data["content"]["parts"]) == 1 + assert data["content"]["parts"][0]["text"] == "Hello, world!" + assert len(response.data) == 1 + diff --git a/tests/test_litellm/proxy/_experimental/mcp_server/test_mcp_sigv4_auth.py b/tests/test_litellm/proxy/_experimental/mcp_server/test_mcp_sigv4_auth.py index 715bb8e8ae..a2295e1271 100644 --- a/tests/test_litellm/proxy/_experimental/mcp_server/test_mcp_sigv4_auth.py +++ b/tests/test_litellm/proxy/_experimental/mcp_server/test_mcp_sigv4_auth.py @@ -2,11 +2,14 @@ Tests for AWS SigV4 authentication in MCP client. Tests the MCPSigV4Auth httpx.Auth subclass that enables per-request -SigV4 signing for Bedrock AgentCore MCP servers. +SigV4 signing for Bedrock AgentCore MCP servers, plus DB/UI path +tests for credential encryption, merge-on-update, and build_from_table. """ +import json + import pytest -from unittest.mock import patch, MagicMock +from unittest.mock import patch, MagicMock, AsyncMock import httpx @@ -315,3 +318,568 @@ class TestMCPServerManagerSigV4: client = await manager._create_mcp_client(server=server) assert client._aws_auth is None + + +class TestSigV4CredentialEncryption: + """Test encrypt/decrypt round-trip for AWS SigV4 credentials.""" + + def test_encrypt_credentials_handles_aws_fields(self): + """AWS credential fields are encrypted in the credentials dict.""" + from litellm.proxy._experimental.mcp_server.db import encrypt_credentials + + creds = { + "aws_access_key_id": "AKIAIOSFODNN7EXAMPLE", + "aws_secret_access_key": "wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY", + "aws_session_token": "FwoGZX...", + "aws_region_name": "us-east-1", + "aws_service_name": "bedrock-agentcore", + } + + with patch( + "litellm.proxy._experimental.mcp_server.db.encrypt_value_helper", + side_effect=lambda value, new_encryption_key: f"enc:{value}", + ): + result = encrypt_credentials(credentials=creds, encryption_key="test-key") + + # Secrets should be encrypted + assert result["aws_access_key_id"] == "enc:AKIAIOSFODNN7EXAMPLE" + assert ( + result["aws_secret_access_key"] + == "enc:wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY" + ) + assert result["aws_session_token"] == "enc:FwoGZX..." + # Non-secrets should be unchanged + assert result["aws_region_name"] == "us-east-1" + assert result["aws_service_name"] == "bedrock-agentcore" + + def test_encrypt_credentials_skips_absent_aws_fields(self): + """encrypt_credentials does not fail when AWS fields are absent.""" + from litellm.proxy._experimental.mcp_server.db import encrypt_credentials + + creds = {"auth_value": "some-token"} + + with patch( + "litellm.proxy._experimental.mcp_server.db.encrypt_value_helper", + side_effect=lambda value, new_encryption_key: f"enc:{value}", + ): + result = encrypt_credentials(credentials=creds, encryption_key="test-key") + + assert result["auth_value"] == "enc:some-token" + assert "aws_access_key_id" not in result + + +class TestCredentialMergeOnUpdate: + """Test that partial credential updates preserve existing fields.""" + + @pytest.mark.asyncio + async def test_partial_update_preserves_existing_credentials(self): + """Updating only aws_region_name should not wipe aws_secret_access_key.""" + from litellm.proxy._experimental.mcp_server.db import update_mcp_server + from litellm.proxy._types import UpdateMCPServerRequest + + existing_record = MagicMock() + existing_record.auth_type = "aws_sigv4" + existing_record.credentials = json.dumps( + { + "aws_access_key_id": "enc:AKI", + "aws_secret_access_key": "enc:SAK", + "aws_region_name": "us-east-1", + } + ) + + mock_prisma = MagicMock() + mock_prisma.db.litellm_mcpservertable.find_unique = AsyncMock( + return_value=existing_record + ) + mock_prisma.db.litellm_mcpservertable.update = AsyncMock( + return_value=MagicMock() + ) + + data = UpdateMCPServerRequest( + server_id="test-server", + auth_type="aws_sigv4", + credentials={"aws_region_name": "eu-west-1"}, + ) + + with patch( + "litellm.proxy._experimental.mcp_server.db._get_salt_key", + return_value=None, + ), patch( + "litellm.proxy._experimental.mcp_server.db.encrypt_value_helper", + side_effect=lambda value, new_encryption_key: value, + ): + await update_mcp_server(mock_prisma, data, "test-user") + + # Grab the data dict passed to prisma update + update_call = mock_prisma.db.litellm_mcpservertable.update + assert update_call.called + data_dict = update_call.call_args[1]["data"] + merged_creds = json.loads(data_dict["credentials"]) + + # Existing encrypted secrets should be preserved + assert merged_creds["aws_access_key_id"] == "enc:AKI" + assert merged_creds["aws_secret_access_key"] == "enc:SAK" + # New region value should be updated + assert merged_creds["aws_region_name"] == "eu-west-1" + + @pytest.mark.asyncio + async def test_update_without_credentials_preserves_all(self): + """Update with no credentials field should not touch existing credentials.""" + from litellm.proxy._experimental.mcp_server.db import update_mcp_server + from litellm.proxy._types import UpdateMCPServerRequest + + mock_prisma = MagicMock() + mock_prisma.db.litellm_mcpservertable.update = AsyncMock( + return_value=MagicMock() + ) + + data = UpdateMCPServerRequest( + server_id="test-server", + description="Updated description", + ) + + with patch( + "litellm.proxy._experimental.mcp_server.db._get_salt_key", + return_value=None, + ): + await update_mcp_server(mock_prisma, data, "test-user") + + data_dict = mock_prisma.db.litellm_mcpservertable.update.call_args[1]["data"] + assert "credentials" not in data_dict + + @pytest.mark.asyncio + async def test_update_new_server_no_merge(self): + """Update with credentials on a server that has no existing credentials.""" + from litellm.proxy._experimental.mcp_server.db import update_mcp_server + from litellm.proxy._types import UpdateMCPServerRequest + + existing_record = MagicMock() + existing_record.auth_type = "aws_sigv4" + existing_record.credentials = None + + mock_prisma = MagicMock() + mock_prisma.db.litellm_mcpservertable.find_unique = AsyncMock( + return_value=existing_record + ) + mock_prisma.db.litellm_mcpservertable.update = AsyncMock( + return_value=MagicMock() + ) + + data = UpdateMCPServerRequest( + server_id="test-server", + auth_type="aws_sigv4", + credentials={"aws_region_name": "us-east-1"}, + ) + + with patch( + "litellm.proxy._experimental.mcp_server.db._get_salt_key", + return_value=None, + ), patch( + "litellm.proxy._experimental.mcp_server.db.encrypt_value_helper", + side_effect=lambda value, new_encryption_key: value, + ): + await update_mcp_server(mock_prisma, data, "test-user") + + data_dict = mock_prisma.db.litellm_mcpservertable.update.call_args[1]["data"] + stored_creds = json.loads(data_dict["credentials"]) + assert stored_creds == {"aws_region_name": "us-east-1"} + + @pytest.mark.asyncio + async def test_auth_type_change_replaces_credentials_entirely(self): + """Switching auth_type should replace credentials, not merge.""" + from litellm.proxy._experimental.mcp_server.db import update_mcp_server + from litellm.proxy._types import UpdateMCPServerRequest + + existing_record = MagicMock() + existing_record.auth_type = "aws_sigv4" + existing_record.credentials = json.dumps( + { + "aws_access_key_id": "enc:AKI", + "aws_secret_access_key": "enc:SAK", + "aws_region_name": "us-east-1", + } + ) + + mock_prisma = MagicMock() + mock_prisma.db.litellm_mcpservertable.find_unique = AsyncMock( + return_value=existing_record + ) + mock_prisma.db.litellm_mcpservertable.update = AsyncMock( + return_value=MagicMock() + ) + + data = UpdateMCPServerRequest( + server_id="test-server", + auth_type="api_key", + credentials={"auth_value": "my-key"}, + ) + + with patch( + "litellm.proxy._experimental.mcp_server.db._get_salt_key", + return_value=None, + ), patch( + "litellm.proxy._experimental.mcp_server.db.encrypt_value_helper", + side_effect=lambda value, new_encryption_key: f"enc:{value}", + ): + await update_mcp_server(mock_prisma, data, "test-user") + + data_dict = mock_prisma.db.litellm_mcpservertable.update.call_args[1]["data"] + stored_creds = json.loads(data_dict["credentials"]) + # Should only have the new api_key credential, no stale aws_* fields + assert stored_creds == {"auth_value": "enc:my-key"} + + @pytest.mark.asyncio + async def test_same_auth_type_merges_credentials(self): + """Same auth_type should merge credentials (preserve untouched fields).""" + from litellm.proxy._experimental.mcp_server.db import update_mcp_server + from litellm.proxy._types import UpdateMCPServerRequest + + existing_record = MagicMock() + existing_record.auth_type = "oauth2" + existing_record.credentials = json.dumps( + { + "client_id": "enc:id", + "client_secret": "enc:secret", + "scopes": ["read"], + } + ) + + mock_prisma = MagicMock() + mock_prisma.db.litellm_mcpservertable.find_unique = AsyncMock( + return_value=existing_record + ) + mock_prisma.db.litellm_mcpservertable.update = AsyncMock( + return_value=MagicMock() + ) + + data = UpdateMCPServerRequest( + server_id="test-server", + auth_type="oauth2", + credentials={"scopes": ["read", "write"]}, + ) + + with patch( + "litellm.proxy._experimental.mcp_server.db._get_salt_key", + return_value=None, + ), patch( + "litellm.proxy._experimental.mcp_server.db.encrypt_value_helper", + side_effect=lambda value, new_encryption_key: value, + ): + await update_mcp_server(mock_prisma, data, "test-user") + + data_dict = mock_prisma.db.litellm_mcpservertable.update.call_args[1]["data"] + merged_creds = json.loads(data_dict["credentials"]) + assert merged_creds["client_id"] == "enc:id" + assert merged_creds["client_secret"] == "enc:secret" + assert merged_creds["scopes"] == ["read", "write"] + + +class TestSigV4BuildFromTable: + """Test build_mcp_server_from_table correctly loads AWS SigV4 credentials.""" + + @pytest.mark.asyncio + async def test_build_mcp_server_from_table_with_sigv4_credentials(self): + """SigV4 credentials from DB are decrypted and mapped to MCPServer fields.""" + from litellm.proxy._experimental.mcp_server.mcp_server_manager import ( + MCPServerManager, + ) + + table_record = MagicMock() + table_record.server_id = "test-sigv4-server" + table_record.server_name = "sigv4_server" + table_record.alias = None + table_record.description = None + table_record.url = "https://bedrock-agentcore.us-east-1.amazonaws.com/invocations" + table_record.spec_path = None + table_record.transport = "http" + table_record.auth_type = "aws_sigv4" + table_record.mcp_info = {"server_name": "sigv4_server"} + table_record.credentials = json.dumps( + { + "aws_access_key_id": "enc:AKIAEXAMPLE", + "aws_secret_access_key": "enc:SECRET", + "aws_session_token": "enc:TOKEN", + "aws_region_name": "us-east-1", + "aws_service_name": "bedrock-agentcore", + } + ) + table_record.extra_headers = None + table_record.static_headers = None + table_record.command = None + table_record.args = [] + table_record.env = None + table_record.mcp_access_groups = [] + table_record.allowed_tools = [] + table_record.disallowed_tools = None + table_record.allow_all_keys = False + table_record.available_on_public_internet = True + table_record.authorization_url = None + table_record.token_url = None + table_record.registration_url = None + table_record.created_at = None + table_record.updated_at = None + table_record.client_id = None + table_record.client_secret = None + table_record.tool_name_to_display_name = None + table_record.tool_name_to_description = None + table_record.byok_api_key_help_url = None + table_record.oauth2_flow = None + + manager = MCPServerManager() + + with patch( + "litellm.proxy._experimental.mcp_server.mcp_server_manager.decrypt_value_helper", + side_effect=lambda value, key, exception_type, return_original_value: value.replace( + "enc:", "" + ), + ): + server = await manager.build_mcp_server_from_table(table_record) + + assert server.auth_type == "aws_sigv4" + assert server.aws_access_key_id == "AKIAEXAMPLE" + assert server.aws_secret_access_key == "SECRET" + assert server.aws_session_token == "TOKEN" + assert server.aws_region_name == "us-east-1" + assert server.aws_service_name == "bedrock-agentcore" + + @pytest.mark.asyncio + async def test_build_mcp_server_from_table_without_sigv4_credentials(self): + """Non-SigV4 servers still work — AWS fields default to None.""" + from litellm.proxy._experimental.mcp_server.mcp_server_manager import ( + MCPServerManager, + ) + + table_record = MagicMock() + table_record.server_id = "test-bearer-server" + table_record.server_name = "bearer_server" + table_record.alias = None + table_record.description = None + table_record.url = "https://example.com/mcp" + table_record.spec_path = None + table_record.transport = "http" + table_record.auth_type = "bearer_token" + table_record.mcp_info = {"server_name": "bearer_server"} + table_record.credentials = json.dumps({"auth_value": "enc:tok"}) + table_record.extra_headers = None + table_record.static_headers = None + table_record.command = None + table_record.args = [] + table_record.env = None + table_record.mcp_access_groups = [] + table_record.allowed_tools = [] + table_record.disallowed_tools = None + table_record.allow_all_keys = False + table_record.available_on_public_internet = True + table_record.authorization_url = None + table_record.token_url = None + table_record.registration_url = None + table_record.created_at = None + table_record.updated_at = None + table_record.client_id = None + table_record.client_secret = None + table_record.tool_name_to_display_name = None + table_record.tool_name_to_description = None + table_record.byok_api_key_help_url = None + table_record.oauth2_flow = None + + manager = MCPServerManager() + + with patch( + "litellm.proxy._experimental.mcp_server.mcp_server_manager.decrypt_value_helper", + side_effect=lambda value, key, exception_type, return_original_value: value.replace( + "enc:", "" + ), + ): + server = await manager.build_mcp_server_from_table(table_record) + + assert server.auth_type == "bearer_token" + assert server.aws_access_key_id is None + assert server.aws_secret_access_key is None + assert server.aws_session_token is None + assert server.aws_region_name is None + assert server.aws_service_name is None + + +class TestDecryptCredentials: + """Test decrypt_credentials helper.""" + + def test_decrypt_credentials_handles_all_secret_fields(self): + """All secret fields are decrypted; non-secret fields are left as-is.""" + from litellm.proxy._experimental.mcp_server.db import decrypt_credentials + + creds = { + "auth_value": "enc:tok", + "client_id": "enc:cid", + "client_secret": "enc:csec", + "aws_access_key_id": "enc:AKI", + "aws_secret_access_key": "enc:SAK", + "aws_session_token": "enc:TOK", + "aws_region_name": "us-east-1", + "aws_service_name": "bedrock-agentcore", + } + + with patch( + "litellm.proxy._experimental.mcp_server.db.decrypt_value_helper", + side_effect=lambda value, key, exception_type="error", return_original_value=False: value.replace("enc:", ""), + ): + result = decrypt_credentials(credentials=creds) + + assert result["auth_value"] == "tok" + assert result["client_id"] == "cid" + assert result["client_secret"] == "csec" + assert result["aws_access_key_id"] == "AKI" + assert result["aws_secret_access_key"] == "SAK" + assert result["aws_session_token"] == "TOK" + # Non-secrets untouched + assert result["aws_region_name"] == "us-east-1" + assert result["aws_service_name"] == "bedrock-agentcore" + + def test_decrypt_credentials_skips_absent_fields(self): + """Absent fields are not touched.""" + from litellm.proxy._experimental.mcp_server.db import decrypt_credentials + + creds = {"aws_access_key_id": "enc:AKI"} + + with patch( + "litellm.proxy._experimental.mcp_server.db.decrypt_value_helper", + side_effect=lambda value, key, exception_type="error", return_original_value=False: value.replace("enc:", ""), + ): + result = decrypt_credentials(credentials=creds) + + assert result["aws_access_key_id"] == "AKI" + assert "aws_secret_access_key" not in result + + +class TestRotateCredentials: + """Test rotate_mcp_server_credentials_master_key decrypts before re-encrypting.""" + + @pytest.mark.asyncio + async def test_rotation_decrypts_then_reencrypts(self): + """Key rotation should decrypt with old key then encrypt with new key.""" + from litellm.proxy._experimental.mcp_server.db import ( + rotate_mcp_server_credentials_master_key, + ) + + server = MagicMock() + server.server_id = "srv-1" + server.credentials = { + "aws_access_key_id": "enc_old:AKI", + "aws_secret_access_key": "enc_old:SAK", + "aws_region_name": "us-east-1", + } + + mock_prisma = MagicMock() + mock_prisma.db.litellm_mcpservertable.find_many = AsyncMock( + return_value=[server] + ) + mock_prisma.db.litellm_mcpservertable.update = AsyncMock() + + with patch( + "litellm.proxy._experimental.mcp_server.db._get_salt_key", + return_value="old-key", + ), patch( + "litellm.proxy._experimental.mcp_server.db.decrypt_value_helper", + side_effect=lambda value, key, exception_type="error", return_original_value=False: value.replace("enc_old:", ""), + ), patch( + "litellm.proxy._experimental.mcp_server.db.encrypt_value_helper", + side_effect=lambda value, new_encryption_key: f"enc_new:{value}", + ): + await rotate_mcp_server_credentials_master_key( + mock_prisma, "admin", "new-key" + ) + + update_call = mock_prisma.db.litellm_mcpservertable.update + assert update_call.called + stored_creds = json.loads(update_call.call_args[1]["data"]["credentials"]) + # Should be decrypted from old, then encrypted with new + assert stored_creds["aws_access_key_id"] == "enc_new:AKI" + assert stored_creds["aws_secret_access_key"] == "enc_new:SAK" + # Non-secret fields should pass through unchanged + assert stored_creds["aws_region_name"] == "us-east-1" + + +class TestAuthTypeSwitchClearsCredentials: + """Test that switching auth_type without credentials clears stale secrets.""" + + @pytest.mark.asyncio + async def test_auth_type_change_without_credentials_clears_stale(self): + """Changing auth_type without providing credentials should clear old ones.""" + from litellm.proxy._experimental.mcp_server.db import update_mcp_server + from litellm.proxy._types import UpdateMCPServerRequest + + existing_record = MagicMock() + existing_record.auth_type = "oauth2" + existing_record.credentials = json.dumps( + {"client_id": "enc:cid", "client_secret": "enc:csec"} + ) + + mock_prisma = MagicMock() + mock_prisma.db.litellm_mcpservertable.find_unique = AsyncMock( + return_value=existing_record + ) + mock_prisma.db.litellm_mcpservertable.update = AsyncMock( + return_value=MagicMock() + ) + + data = UpdateMCPServerRequest( + server_id="test-server", + auth_type="aws_sigv4", + # No credentials provided + ) + + with patch( + "litellm.proxy._experimental.mcp_server.db._get_salt_key", + return_value=None, + ): + await update_mcp_server(mock_prisma, data, "test-user") + + data_dict = mock_prisma.db.litellm_mcpservertable.update.call_args[1]["data"] + # Credentials should be cleared (set to None) + assert data_dict.get("credentials") is None + + +class TestInheritCredentials: + """Test _inherit_credentials_from_existing_server copies AWS fields.""" + + def test_inherits_sigv4_credentials(self): + """SigV4 fields are copied from existing server to inherited credentials.""" + from litellm.proxy.management_endpoints.mcp_management_endpoints import ( + _inherit_credentials_from_existing_server, + ) + from litellm.proxy._types import NewMCPServerRequest + from litellm.types.mcp_server.mcp_server_manager import MCPServer + + existing = MCPServer( + server_id="existing-sigv4", + name="sigv4_server", + server_name="sigv4_server", + url="https://bedrock-agentcore.us-east-1.amazonaws.com/mcp", + transport=MCPTransport.http, + auth_type=MCPAuth.aws_sigv4, + aws_access_key_id="AKIAEXAMPLE", + aws_secret_access_key="SECRET", + aws_session_token="TOKEN", + aws_region_name="us-east-1", + aws_service_name="bedrock-agentcore", + ) + + payload = NewMCPServerRequest( + server_id="existing-sigv4", + server_name="sigv4_server", + url="https://bedrock-agentcore.us-east-1.amazonaws.com/mcp", + transport="http", + auth_type="aws_sigv4", + ) + + with patch( + "litellm.proxy.management_endpoints.mcp_management_endpoints.global_mcp_server_manager" + ) as mock_manager: + mock_manager.get_mcp_server_by_id.return_value = existing + result = _inherit_credentials_from_existing_server(payload) + + assert result.credentials is not None + assert result.credentials["aws_access_key_id"] == "AKIAEXAMPLE" + assert result.credentials["aws_secret_access_key"] == "SECRET" + assert result.credentials["aws_session_token"] == "TOKEN" + assert result.credentials["aws_region_name"] == "us-east-1" + assert result.credentials["aws_service_name"] == "bedrock-agentcore" diff --git a/tests/test_litellm/proxy/pass_through_endpoints/test_pass_through_endpoints.py b/tests/test_litellm/proxy/pass_through_endpoints/test_pass_through_endpoints.py index 7ec97ddc18..5af24f9612 100644 --- a/tests/test_litellm/proxy/pass_through_endpoints/test_pass_through_endpoints.py +++ b/tests/test_litellm/proxy/pass_through_endpoints/test_pass_through_endpoints.py @@ -2369,3 +2369,107 @@ def test_get_registered_pass_through_route_with_custom_root(): # Clean up _registered_pass_through_routes.clear() + + +def test_mapped_pass_through_routes_with_server_root_path(): + """ + Mapped passthrough routes (vertex_ai, bedrock, etc) should match + even when SERVER_ROOT_PATH is set and the incoming route is prefixed. + + Regression test for https://github.com/BerriAI/litellm/issues/22272 + """ + from litellm.proxy.pass_through_endpoints.pass_through_endpoints import ( + InitPassThroughEndpointHelpers, + ) + + with patch( + "litellm.proxy.pass_through_endpoints.pass_through_endpoints.get_server_root_path" + ) as mock_get_root: + mock_get_root.return_value = "/litellm" + + # prefixed route should match mapped routes like /vertex_ai + assert ( + InitPassThroughEndpointHelpers.is_registered_pass_through_route( + "/litellm/vertex_ai/v1/projects/foo" + ) + is True + ) + assert ( + InitPassThroughEndpointHelpers.is_registered_pass_through_route( + "/litellm/bedrock/model/invoke" + ) + is True + ) + + # bare route without prefix should not match when root is set + assert ( + InitPassThroughEndpointHelpers.is_registered_pass_through_route( + "/vertex_ai/v1/projects/foo" + ) + is False + ) + + +@pytest.mark.asyncio +async def test_multipart_passthrough_preserves_boundary(): + """ + Test that multipart/form-data requests through passthrough preserve the boundary + and can be correctly parsed by the upstream server. + + Regression test for multipart boundary stripping issue. + """ + from io import BytesIO + + # Mock the httpx request to verify files are passed correctly + mock_response = MagicMock() + mock_response.status_code = 200 + mock_response.headers = httpx.Headers({"content-type": "application/json"}) + mock_response.aread = AsyncMock(return_value=b'{"filename": "test.txt", "size": 17}') + mock_response.text = '{"filename": "test.txt", "size": 17}' + + async def mock_httpx_request(method, url, **kwargs): + # Verify that files parameter is passed (not json) + assert "files" in kwargs, "Files should be passed for multipart requests" + assert "file" in kwargs["files"], "File field should be in files dict" + + # Verify content-type is NOT in headers (httpx will set it with correct boundary) + headers = kwargs.get("headers", {}) + assert "content-type" not in headers, "content-type should be removed for multipart" + + filename, content, content_type = kwargs["files"]["file"] + assert filename == "test.txt" + assert content == b"test file content" + assert content_type == "text/plain" + + return mock_response + + async_client = MagicMock() + async_client.request = AsyncMock(side_effect=mock_httpx_request) + + # Create mock request + request = MagicMock(spec=Request) + request.method = "POST" + request.headers = Headers({"content-type": "multipart/form-data; boundary=test123"}) + + # Mock form data + file_content = b"test file content" + file = BytesIO(file_content) + headers = Headers({"content-type": "text/plain"}) + upload_file = UploadFile(file=file, filename="test.txt", headers=headers) + upload_file.read = AsyncMock(return_value=file_content) + + form_data = {"file": upload_file} + request.form = AsyncMock(return_value=form_data) + + # Test the multipart handler directly + response = await HttpPassThroughEndpointHelpers.make_multipart_http_request( + request=request, + async_client=async_client, + url=httpx.URL("http://test.com/upload"), + headers={}, + requested_query_params=None, + ) + + # Verify the response + assert response.status_code == 200 + async_client.request.assert_called_once() diff --git a/tests/test_litellm/proxy/test_proxy_cli.py b/tests/test_litellm/proxy/test_proxy_cli.py index cf6511c18a..642d21a42f 100644 --- a/tests/test_litellm/proxy/test_proxy_cli.py +++ b/tests/test_litellm/proxy/test_proxy_cli.py @@ -664,6 +664,64 @@ class TestHealthAppFactory: ) mock_setup_database.assert_called_with(use_migrate=False) + @patch("subprocess.run") + @patch("atexit.register") + @patch("litellm.proxy.db.prisma_client.PrismaManager.setup_database") + @patch("litellm.proxy.db.check_migration.check_prisma_schema_diff") + @patch("litellm.proxy.db.prisma_client.should_update_prisma_schema") + def test_startup_fails_when_db_setup_fails( + self, + mock_should_update_schema, + mock_check_schema_diff, + mock_setup_database, + mock_atexit_register, + mock_subprocess_run, + ): + """Test that proxy exits with code 1 when PrismaManager.setup_database returns False""" + from litellm.proxy.proxy_cli import run_server + + mock_subprocess_run.return_value = MagicMock(returncode=0) + mock_should_update_schema.return_value = True + mock_setup_database.return_value = False + + mock_proxy_module = MagicMock( + app=MagicMock(), + ProxyConfig=MagicMock(), + KeyManagementSettings=MagicMock(), + save_worker_config=MagicMock(), + ) + + clean_env = { + k: v + for k, v in os.environ.items() + if k not in ("DATABASE_URL", "DIRECT_URL") + } + clean_env["DATABASE_URL"] = "postgresql://test:test@localhost:5432/test" + + with patch.dict( + os.environ, clean_env, clear=True + ), patch.dict( + "sys.modules", + { + "proxy_server": mock_proxy_module, + "litellm.proxy.proxy_server": mock_proxy_module, + }, + ), patch( + "litellm.proxy.proxy_cli.ProxyInitializationHelpers._get_default_unvicorn_init_args" + ) as mock_get_args: + mock_get_args.return_value = { + "app": "litellm.proxy.proxy_server:app", + "host": "localhost", + "port": 8000, + } + + with pytest.raises(SystemExit) as exc_info: + run_server.main( + ["--local", "--skip_server_startup"], standalone_mode=False + ) + assert exc_info.value.code == 1 + mock_setup_database.assert_called_once_with(use_migrate=True) + # --- Module-level helpers for worker startup hook tests --- diff --git a/ui/litellm-dashboard/src/components/chat/types.ts b/ui/litellm-dashboard/src/components/chat/types.ts index 4c0da7a095..9b657306db 100644 --- a/ui/litellm-dashboard/src/components/chat/types.ts +++ b/ui/litellm-dashboard/src/components/chat/types.ts @@ -1,4 +1,5 @@ -export type { MCPEvent } from "../mcp_tools/types"; +import type { MCPEvent } from "../mcp_tools/types"; +export type { MCPEvent }; export interface ChatMessage { id: string; diff --git a/ui/litellm-dashboard/src/components/mcp_tools/create_mcp_server.tsx b/ui/litellm-dashboard/src/components/mcp_tools/create_mcp_server.tsx index 74945731a2..e684540289 100644 --- a/ui/litellm-dashboard/src/components/mcp_tools/create_mcp_server.tsx +++ b/ui/litellm-dashboard/src/components/mcp_tools/create_mcp_server.tsx @@ -33,7 +33,7 @@ interface CreateMCPServerProps { } const AUTH_TYPES_REQUIRING_AUTH_VALUE = [AUTH_TYPE.API_KEY, AUTH_TYPE.BEARER_TOKEN, AUTH_TYPE.TOKEN, AUTH_TYPE.BASIC]; -const AUTH_TYPES_REQUIRING_CREDENTIALS = [...AUTH_TYPES_REQUIRING_AUTH_VALUE, AUTH_TYPE.OAUTH2]; +const AUTH_TYPES_REQUIRING_CREDENTIALS = [...AUTH_TYPES_REQUIRING_AUTH_VALUE, AUTH_TYPE.OAUTH2, AUTH_TYPE.AWS_SIGV4]; const CREATE_OAUTH_UI_STATE_KEY = "litellm-mcp-oauth-create-state"; const reduceStaticHeaders = (list: unknown): Record => { @@ -85,6 +85,7 @@ const CreateMCPServer: React.FC = ({ const authType = formValues.auth_type as string | undefined; const shouldShowAuthValueField = authType ? AUTH_TYPES_REQUIRING_AUTH_VALUE.includes(authType) : false; const isOAuthAuthType = authType === AUTH_TYPE.OAUTH2; + const isAwsSigV4AuthType = authType === AUTH_TYPE.AWS_SIGV4; const isM2MFlow = isOAuthAuthType && formValues.oauth_flow_type === OAUTH_FLOW.M2M; const persistCreateUiState = () => { @@ -767,6 +768,7 @@ const CreateMCPServer: React.FC = ({ Token Basic Auth OAuth + AWS SigV4 (Bedrock AgentCore MCPs) @@ -818,6 +820,122 @@ const CreateMCPServer: React.FC = ({ /> )} + {transportType !== "stdio" && transportType !== "" && isAwsSigV4AuthType && ( + <> +

+ For MCP servers hosted on AWS Bedrock AgentCore.{" "} + + View docs → + +

+ + AWS Region + + + + + } + name={["credentials", "aws_region_name"]} + rules={[{ required: true, message: "AWS region is required for SigV4 auth" }]} + > + + + + AWS Service Name + + + + + } + name={["credentials", "aws_service_name"]} + > + + + + AWS Access Key ID + + + + + } + name={["credentials", "aws_access_key_id"]} + dependencies={[["credentials", "aws_secret_access_key"]]} + rules={[ + ({ getFieldValue }) => ({ + validator(_, value) { + const secretKey = getFieldValue(["credentials", "aws_secret_access_key"]); + if (secretKey && !value) { + return Promise.reject(new Error("Access Key ID is required when Secret Access Key is provided")); + } + return Promise.resolve(); + }, + }), + ]} + > + + + + AWS Secret Access Key + + + + + } + name={["credentials", "aws_secret_access_key"]} + dependencies={[["credentials", "aws_access_key_id"]]} + rules={[ + ({ getFieldValue }) => ({ + validator(_, value) { + const accessKeyId = getFieldValue(["credentials", "aws_access_key_id"]); + if (accessKeyId && !value) { + return Promise.reject(new Error("Secret Access Key is required when Access Key ID is provided")); + } + return Promise.resolve(); + }, + }), + ]} + > + + + + AWS Session Token + + + + + } + name={["credentials", "aws_session_token"]} + > + + + + )} + {/* Stdio Configuration - only show for stdio transport */} diff --git a/ui/litellm-dashboard/src/components/mcp_tools/mcp_server_edit.tsx b/ui/litellm-dashboard/src/components/mcp_tools/mcp_server_edit.tsx index eadf93d8a9..04cce34303 100644 --- a/ui/litellm-dashboard/src/components/mcp_tools/mcp_server_edit.tsx +++ b/ui/litellm-dashboard/src/components/mcp_tools/mcp_server_edit.tsx @@ -22,7 +22,7 @@ interface MCPServerEditProps { } const AUTH_TYPES_REQUIRING_AUTH_VALUE = [AUTH_TYPE.API_KEY, AUTH_TYPE.BEARER_TOKEN, AUTH_TYPE.TOKEN, AUTH_TYPE.BASIC]; -const AUTH_TYPES_REQUIRING_CREDENTIALS = [...AUTH_TYPES_REQUIRING_AUTH_VALUE, AUTH_TYPE.OAUTH2]; +const AUTH_TYPES_REQUIRING_CREDENTIALS = [...AUTH_TYPES_REQUIRING_AUTH_VALUE, AUTH_TYPE.OAUTH2, AUTH_TYPE.AWS_SIGV4]; const EDIT_OAUTH_UI_STATE_KEY = "litellm-mcp-oauth-edit-state"; const MCPServerEdit: React.FC = ({ @@ -50,6 +50,7 @@ const MCPServerEdit: React.FC = ({ const isMCPTransport = !isStdioTransport && !isOpenAPITransport; const shouldShowAuthValueField = authType ? AUTH_TYPES_REQUIRING_AUTH_VALUE.includes(authType) : false; const isOAuthAuthType = authType === AUTH_TYPE.OAUTH2; + const isAwsSigV4AuthType = authType === AUTH_TYPE.AWS_SIGV4; const oauthFlowTypeValue = Form.useWatch("oauth_flow_type", form) as string | undefined; const isM2MFlow = isOAuthAuthType && oauthFlowTypeValue === OAUTH_FLOW.M2M; @@ -665,6 +666,7 @@ const MCPServerEdit: React.FC = ({ Token Basic Auth OAuth + AWS SigV4 (Bedrock AgentCore MCPs) )} @@ -883,6 +885,100 @@ const MCPServerEdit: React.FC = ({ )} + {!isStdioTransport && isAwsSigV4AuthType && ( + <> +

+ For MCP servers hosted on AWS Bedrock AgentCore.{" "} + + View docs → + +

+ + AWS Region + + + + + } + name={["credentials", "aws_region_name"]} + rules={[]} + > + + + + AWS Service Name + + + + + } + name={["credentials", "aws_service_name"]} + > + + + + AWS Access Key ID + + + + + } + name={["credentials", "aws_access_key_id"]} + rules={[]} + > + + + + AWS Secret Access Key + + + + + } + name={["credentials", "aws_secret_access_key"]} + rules={[]} + > + + + + AWS Session Token + + + + + } + name={["credentials", "aws_session_token"]} + > + + + + )} + {/* Permission Management / Access Control Section */}