fix(messages): honor chat completions opt-in

litellm/llms/anthropic/experimental_pass_through/messages/handler.py

@@ -42,13 +42,16 @@ from .utils import AnthropicMessagesRequestUtils, mock_response
 _RESPONSES_API_PROVIDERS: Final = frozenset({"openai"})
 
 
-def _should_route_to_responses_api(custom_llm_provider: str | None) -> bool:
+def _should_route_to_responses_api(
+    custom_llm_provider: str | None,
+    use_chat_completions_api: bool | None = None,
+) -> bool:
     """Return True when the provider should use the Responses API path.
 
     Set ``litellm.use_chat_completions_url_for_anthropic_messages = True`` to
     opt out and route OpenAI/Azure requests through chat/completions instead.
     """
-    if litellm.use_chat_completions_url_for_anthropic_messages:
+    if litellm.use_chat_completions_url_for_anthropic_messages or use_chat_completions_api is True:
         return False
     return custom_llm_provider in _RESPONSES_API_PROVIDERS
 
@@ -551,7 +554,10 @@ def anthropic_messages_handler(
             custom_llm_provider=custom_llm_provider,
             **kwargs,
         )
-        if _should_route_to_responses_api(custom_llm_provider):
+        if _should_route_to_responses_api(
+            custom_llm_provider=custom_llm_provider,
+            use_chat_completions_api=kwargs.get("use_chat_completions_api"),
+        ):
             return LiteLLMMessagesToResponsesAPIHandler.anthropic_messages_handler(**_shared_kwargs)
 
         # The in-gateway context_management polyfill runs inside

tests/test_litellm/llms/anthropic/experimental_pass_through/messages/test_anthropic_experimental_pass_through_messages_handler.py

@@ -939,6 +939,28 @@ def test_gate_translates_when_supported_endpoints_absent(monkeypatch):
     assert "config" not in captured
 
 
+def test_gate_uses_chat_completions_when_requested(monkeypatch):
+    """The per-deployment chat-completions opt-in must bypass Responses API."""
+    from litellm.llms.anthropic.experimental_pass_through.messages.handler import (
+        anthropic_messages_handler,
+    )
+
+    captured, translation_calls = _gate_stubs(monkeypatch)
+
+    result = anthropic_messages_handler(
+        max_tokens=100,
+        messages=[{"role": "user", "content": "Hello"}],
+        model="openai/some-model",
+        api_key="sk-test",
+        api_base="https://host/v1",
+        use_chat_completions_api=True,
+    )
+
+    assert result == "translated"
+    assert translation_calls["count"] == 1
+    assert "config" not in captured
+
+
 def test_gate_passthrough_skipped_when_only_chat_completions_supported(monkeypatch):
     """A deployment that lists only /v1/chat/completions is still translated;
     the opt-in is specifically the /v1/messages entry."""