fix(vertex_ai): forward extra_body to completion transformation handler (#20950)

* fix(vertex_ai): forward extra_body to completion transformation handler

The responses() function accepted extra_body as a named parameter but
did not pass it to response_api_handler when responses_api_provider_config
was None (completion transformation path), silently dropping it.

Also adds deep-merge support for extra_body in Vertex AI Gemini
transformation, so dict values like generationConfig are merged rather
than replaced.

* refactor(vertex_ai): extract _merge_extra_body to fix PLR0915 lint

Move the extra_body merge loop into a helper function to keep
_transform_request_body under the 50-statement limit.

litellm/llms/vertex_ai/gemini/transformation.py

@@ -529,6 +529,17 @@ def _gemini_convert_messages_with_history(  # noqa: PLR0915
         raise e
 
 
+def _pop_and_merge_extra_body(data: RequestBody, optional_params: dict) -> None:
+    """Pop extra_body from optional_params and shallow-merge into data, deep-merging dict values."""
+    extra_body: Optional[dict] = optional_params.pop("extra_body", None)
+    if extra_body is not None:
+        for k, v in extra_body.items():
+            if k in data and isinstance(data[k], dict) and isinstance(v, dict):
+                data[k].update(v)
+            else:
+                data[k] = v
+
+
 def _transform_request_body(
     messages: List[AllMessageValues],
     model: str,
@@ -619,6 +630,7 @@ def _transform_request_body(
         # Only add labels for Vertex AI endpoints (not Google GenAI/AI Studio) and only if non-empty
         if labels and custom_llm_provider != LlmProviders.GEMINI:
             data["labels"] = labels
+        _pop_and_merge_extra_body(data, optional_params)
     except Exception as e:
         raise e
 

litellm/responses/main.py

@@ -682,6 +682,7 @@ def responses(
                 _is_async=_is_async,
                 stream=stream,
                 extra_headers=extra_headers,
+                extra_body=extra_body,
                 **kwargs,
             )
 

tests/test_litellm/responses/test_responses_utils.py

@@ -2,6 +2,7 @@ import base64
 import json
 import os
 import sys
+from unittest.mock import MagicMock, patch
 
 import pytest
 from fastapi.testclient import TestClient
@@ -352,3 +353,34 @@ class TestResponsesAPIProviderSpecificParams:
         # Should not raise any exception
         result = ResponsesAPIRequestUtils.get_requested_response_api_optional_param(params)
         assert "temperature" in result
+
+
+def test_responses_extra_body_forwarded_to_completion_transformation_handler():
+    """
+    Regression test: extra_body must be forwarded to response_api_handler
+    when responses_api_provider_config is None (completion transformation path).
+
+    Before the fix, extra_body was a named parameter of responses() but was
+    not passed to litellm_completion_transformation_handler.response_api_handler(),
+    so it was silently dropped.
+    """
+    with patch(
+        "litellm.responses.main.ProviderConfigManager.get_provider_responses_api_config",
+        return_value=None,
+    ), patch(
+        "litellm.responses.main.litellm_completion_transformation_handler.response_api_handler",
+    ) as mock_handler:
+        mock_handler.return_value = MagicMock()
+
+        litellm.responses(
+            model="openai/gpt-4o",
+            input="Hello",
+            extra_body={"custom_key": "custom_value"},
+        )
+
+        mock_handler.assert_called_once()
+        call_kwargs = mock_handler.call_args
+        # extra_body can be a positional or keyword arg; check both
+        assert call_kwargs.kwargs.get("extra_body") == {
+            "custom_key": "custom_value"
+        }