From 0d84c11bbead2f39247f727f3ddea5855d7d5e6b Mon Sep 17 00:00:00 2001
From: pale-aura <92485969+pale-aura@users.noreply.github.com>
Date: Thu, 30 Oct 2025 22:40:35 -0400
Subject: [PATCH] Fix vercel_ai_gateway entry for glm-4.6 (#16084)

* Fix vercel_ai_gateway entry for glm-4.6

* Fix: add unit test for vercel_ai_gateway/zai/glm-4.6

* Fix:

Add fix for unit test on vercel_ai_gateway/zai/glm-4.6

* Remove print statements in test_vercel_ai_gateway_glm46_cost_math
---
 model_prices_and_context_window.json          | 24 +++++++++++--------
 .../test_vercel_ai_gateway.py                 | 21 +++++++++++++++-
 2 files changed, 34 insertions(+), 11 deletions(-)

diff --git a/model_prices_and_context_window.json b/model_prices_and_context_window.json
index e685ef6a7b..624dea84e1 100644
--- a/model_prices_and_context_window.json
+++ b/model_prices_and_context_window.json
@@ -21145,16 +21145,6 @@
         "mode": "chat",
         "output_cost_per_token": 2.4e-07
     },
-    "vercel_ai_gateway/glm-4.6": {
-        "litellm_provider": "vercel_ai_gateway",
-        "cache_read_input_token_cost": 1.1e-07,
-        "input_cost_per_token": 6e-07,
-        "max_input_tokens": 200000,
-        "max_output_tokens": 200000,
-        "max_tokens": 200000,
-        "mode": "chat",
-        "output_cost_per_token": 2.2e-06
-    },
     "vercel_ai_gateway/alibaba/qwen-3-235b": {
         "input_cost_per_token": 2e-07,
         "litellm_provider": "vercel_ai_gateway",
@@ -21988,6 +21978,20 @@
         "mode": "chat",
         "output_cost_per_token": 1.1e-06
     },
+    "vercel_ai_gateway/zai/glm-4.6": {
+        "litellm_provider": "vercel_ai_gateway",
+        "cache_read_input_token_cost": 1.1e-07,
+        "input_cost_per_token": 4.5e-07,
+        "max_input_tokens": 200000,
+        "max_output_tokens": 200000,
+        "max_tokens": 200000,
+        "mode": "chat",
+        "output_cost_per_token": 1.8e-06,
+        "source": "https://vercel.com/ai-gateway/models/glm-4.6",
+        "supports_function_calling": true,
+        "supports_parallel_function_calling": true,
+        "supports_tool_choice": true
+    },
     "vertex_ai/claude-3-5-haiku": {
         "input_cost_per_token": 1e-06,
         "litellm_provider": "vertex_ai-anthropic_models",
diff --git a/tests/test_litellm/llms/vercel_ai_gateway/test_vercel_ai_gateway.py b/tests/test_litellm/llms/vercel_ai_gateway/test_vercel_ai_gateway.py
index 7a2d992ec8..f4d4730e84 100755
--- a/tests/test_litellm/llms/vercel_ai_gateway/test_vercel_ai_gateway.py
+++ b/tests/test_litellm/llms/vercel_ai_gateway/test_vercel_ai_gateway.py
@@ -10,7 +10,8 @@ import respx
 import litellm
 from litellm import completion
 from litellm.llms.vercel_ai_gateway.chat.transformation import VercelAIGatewayConfig
-
+from litellm.cost_calculator import cost_per_token
+import math
 
 @pytest.fixture
 def vercel_ai_gateway_response():
@@ -226,3 +227,21 @@ def test_vercel_ai_gateway_models_endpoint_failure():
 
         with pytest.raises(Exception, match="Failed to get models: Not found"):
             config.get_models()
+
+def test_vercel_ai_gateway_glm46_cost_math():
+    """Test the cost math for glm-4.6"""
+
+    with open("model_prices_and_context_window.json", "r") as f:
+        litellm.model_cost = json.load(f)
+
+    key = "vercel_ai_gateway/zai/glm-4.6"
+    info = litellm.model_cost[key]
+
+    prompt_cost, completion_cost = cost_per_token(
+        model="vercel_ai_gateway/zai/glm-4.6",
+        prompt_tokens=1000,
+        completion_tokens=500,
+    )
+
+    assert math.isclose(prompt_cost, 1000 * info["input_cost_per_token"], rel_tol=1e-12)
+    assert math.isclose(completion_cost, 500 * info["output_cost_per_token"], rel_tol=1e-12)