diff --git a/litellm/llms/openai/vector_stores/transformation.py b/litellm/llms/openai/vector_stores/transformation.py
index 0e890f0fd5..76cd12be8e 100644
--- a/litellm/llms/openai/vector_stores/transformation.py
+++ b/litellm/llms/openai/vector_stores/transformation.py
@@ -14,6 +14,7 @@ from litellm.types.vector_stores import (
     VectorStoreSearchRequest,
     VectorStoreSearchResponse,
 )
+from litellm.utils import add_openai_metadata
 
 if TYPE_CHECKING:
     from litellm.litellm_core_utils.litellm_logging import Logging as _LiteLLMLoggingObj
@@ -119,12 +120,13 @@ class OpenAIVectorStoreConfig(BaseVectorStoreConfig):
         api_base: str,
     ) -> Tuple[str, Dict]:
         url = api_base  # Base URL for creating vector stores
+        metadata = vector_store_create_optional_params.get("metadata", None)
         typed_request_body = VectorStoreCreateRequest(
             name=vector_store_create_optional_params.get("name", None),
             file_ids=vector_store_create_optional_params.get("file_ids", None),
             expires_after=vector_store_create_optional_params.get("expires_after", None),
             chunking_strategy=vector_store_create_optional_params.get("chunking_strategy", None),
-            metadata=vector_store_create_optional_params.get("metadata", None),
+            metadata=add_openai_metadata(metadata) if metadata is not None else None,
         )
 
         dict_request_body = cast(dict, typed_request_body)
diff --git a/litellm/proxy/route_llm_request.py b/litellm/proxy/route_llm_request.py
index e57539a612..eb48f7de2d 100644
--- a/litellm/proxy/route_llm_request.py
+++ b/litellm/proxy/route_llm_request.py
@@ -153,6 +153,8 @@ async def route_request(
                 "aget_responses",
                 "adelete_responses",
                 "alist_input_items",
+                "avector_store_create",
+                "avector_store_search",
             ]:
                 # moderation endpoint does not require `model` parameter
                 return getattr(llm_router, f"{route_type}")(**data)
diff --git a/tests/test_litellm/llms/openai/vector_stores/test_openai_vector_stores_transformation.py b/tests/test_litellm/llms/openai/vector_stores/test_openai_vector_stores_transformation.py
new file mode 100644
index 0000000000..17a611f571
--- /dev/null
+++ b/tests/test_litellm/llms/openai/vector_stores/test_openai_vector_stores_transformation.py
@@ -0,0 +1,67 @@
+import pytest
+
+from litellm.llms.openai.vector_stores.transformation import OpenAIVectorStoreConfig
+from litellm.types.vector_stores import (
+    VectorStoreCreateOptionalRequestParams,
+)
+
+
+class TestOpenAIVectorStoreAPIConfig:
+
+    @pytest.mark.parametrize(
+        "metadata", [{}, None]
+    )
+    def test_transform_create_vector_store_request_with_metadata_empty_or_none(self, metadata):
+        """
+        Test transform_create_vector_store_request when metadata is None or empty dict.
+        """
+        config = OpenAIVectorStoreConfig()
+        api_base = "https://api.openai.com/v1/vector_stores"
+        
+        vector_store_create_params: VectorStoreCreateOptionalRequestParams = {
+            "name": "test-vector-store",
+            "file_ids": ["file-123", "file-456"],
+            "metadata": metadata,
+        }
+        
+        url, request_body = config.transform_create_vector_store_request(
+            vector_store_create_params, api_base
+        )
+
+        assert url == api_base
+        assert request_body["name"] == "test-vector-store"
+        assert request_body["file_ids"] == ["file-123", "file-456"]
+        assert request_body["metadata"] == metadata
+
+
+    def test_transform_create_vector_store_request_with_large_metadata(self):
+        """
+        Test transform_create_vector_store_request with metadata exceeding 16 keys.
+        
+        OpenAI limits metadata to 16 keys maximum.
+        """
+        config = OpenAIVectorStoreConfig()
+        api_base = "https://api.openai.com/v1/vector_stores"
+        
+        # Create metadata with more than 16 keys
+        large_metadata = {f"key_{i}": f"value_{i}" for i in range(20)}
+        
+        vector_store_create_params: VectorStoreCreateOptionalRequestParams = {
+            "name": "test-vector-store",
+            "metadata": large_metadata,
+        }
+        
+        url, request_body = config.transform_create_vector_store_request(
+            vector_store_create_params, api_base
+        )
+        
+        assert url == api_base
+        assert request_body["name"] == "test-vector-store"
+        
+        # Should be trimmed to 16 keys
+        assert len(request_body["metadata"]) == 16
+        
+        # Should contain the first 16 keys (as per add_openai_metadata implementation)
+        for i in range(16):
+            assert f"key_{i}" in request_body["metadata"]
+            assert request_body["metadata"][f"key_{i}"] == f"value_{i}"
diff --git a/tests/test_litellm/proxy/test_route_llm_request.py b/tests/test_litellm/proxy/test_route_llm_request.py
index d4a65d8847..9d8aebd2d1 100644
--- a/tests/test_litellm/proxy/test_route_llm_request.py
+++ b/tests/test_litellm/proxy/test_route_llm_request.py
@@ -49,7 +49,7 @@ async def test_route_request_dynamic_credentials(route_type):
 @pytest.mark.asyncio
 async def test_route_request_no_model_required():
     """Test route types that don't require model parameter"""
-    test_cases = ["amoderation", "aget_responses", "adelete_responses"]
+    test_cases = ["amoderation", "aget_responses", "adelete_responses", "avector_store_create", "avector_store_search"]
 
     for route_type in test_cases:
         # Test data without model parameter
@@ -72,7 +72,7 @@ async def test_route_request_no_model_required():
 @pytest.mark.asyncio
 async def test_route_request_no_model_required_with_router_settings():
     """Test route types that don't require model parameter with router settings"""
-    test_cases = ["amoderation", "aget_responses", "adelete_responses"]
+    test_cases = ["amoderation", "aget_responses", "adelete_responses", "avector_store_create", "avector_store_search"]
 
     for route_type in test_cases:
         # Test data with model parameter (it will be ignored for these route types)