test_o1_parallel_tool_calls

litellm/llms/azure/responses/transformation.py

@@ -1,5 +1,5 @@
-from typing import TYPE_CHECKING, Any, Dict, List, Literal, Optional, Tuple, Union
 from copy import deepcopy
+from typing import TYPE_CHECKING, Any, Dict, List, Literal, Optional, Tuple, Union
 
 import httpx
 from openai.types.responses import ResponseReasoningItem
@@ -21,10 +21,25 @@ else:
 
 
 class AzureOpenAIResponsesAPIConfig(OpenAIResponsesAPIConfig):
+
+    # Parameters not supported by Azure Responses API
+    AZURE_UNSUPPORTED_PARAMS = ["context_management"]
+
     @property
     def custom_llm_provider(self) -> LlmProviders:
         return LlmProviders.AZURE
 
+    def get_supported_openai_params(self, model: str) -> list:
+        """
+        Azure Responses API does not support context_management (compaction).
+        """
+        base_supported_params = super().get_supported_openai_params(model)
+        return [
+            param
+            for param in base_supported_params
+            if param not in self.AZURE_UNSUPPORTED_PARAMS
+        ]
+
     def validate_environment(
         self, headers: dict, model: str, litellm_params: Optional[GenericLiteLLMParams]
     ) -> dict:

litellm/model_prices_and_context_window_backup.json

@@ -23204,7 +23204,7 @@
         "mode": "chat",
         "output_cost_per_token": 6e-05,
         "supports_function_calling": true,
-        "supports_parallel_function_calling": true,
+        "supports_parallel_function_calling": false,
         "supports_pdf_input": true,
         "supports_prompt_caching": true,
         "supports_reasoning": true,

model_prices_and_context_window.json

@@ -23204,7 +23204,7 @@
         "mode": "chat",
         "output_cost_per_token": 6e-05,
         "supports_function_calling": true,
-        "supports_parallel_function_calling": true,
+        "supports_parallel_function_calling": false,
         "supports_pdf_input": true,
         "supports_prompt_caching": true,
         "supports_reasoning": true,

tests/llm_responses_api_testing/base_responses_api.py

@@ -694,10 +694,14 @@ class BaseResponsesAPITest(ABC):
         """
         base_completion_call_args = self.get_base_completion_call_args()
         model = base_completion_call_args.get("model") or ""
-        # Only run with context_management for OpenAI (OAI) for now
-        if "openai/" not in str(model) and "azure/" not in str(model):
+        # Azure does not support compaction context_management (only clear_tool_results)
+        if "azure/" in str(model):
             pytest.skip(
-                "context_management server-side compaction e2e is only run for OpenAI/Azure"
+                "context_management compaction is not supported on Azure"
+            )
+        if "openai/" not in str(model):
+            pytest.skip(
+                "context_management server-side compaction e2e is only run for OpenAI"
             )
         context_management = [{"type": "compaction", "compact_threshold": 200000}]
         try:

tests/test_litellm/llms/azure/response/test_azure_transformation.py

@@ -318,6 +318,7 @@ class TestAzureResponsesAPIConfig:
     def test_azure_cancel_response_api_response(self):
         """Test Azure cancel response API response transformation"""
         from unittest.mock import Mock
+
         from litellm.types.llms.openai import ResponsesAPIResponse
 
         # Mock response
@@ -492,3 +493,12 @@ class TestAzureResponsesAPIConfig:
         )
 
         assert "tools" not in response_api_params
+
+    def test_azure_responses_api_context_management_unsupported(self):
+        """Test that context_management is not in Azure supported params.
+
+        Azure does not support context_management (compaction). It should be
+        excluded from supported params so it gets dropped.
+        """
+        supported = self.config.get_supported_openai_params(self.model)
+        assert "context_management" not in supported