diff --git a/litellm/llms/chatgpt/responses/transformation.py b/litellm/llms/chatgpt/responses/transformation.py index 26a5a3f150b..762c5136276 100644 --- a/litellm/llms/chatgpt/responses/transformation.py +++ b/litellm/llms/chatgpt/responses/transformation.py @@ -145,6 +145,7 @@ class ChatGPTResponsesAPIConfig(OpenAIResponsesAPIConfig): "tools", "tool_choice", "reasoning", + "prompt_cache_key", "previous_response_id", "truncation", } diff --git a/tests/unit/llms/chatgpt/responses/test_chatgpt_responses_transformation.py b/tests/unit/llms/chatgpt/responses/test_chatgpt_responses_transformation.py index 635363e77ea..4061779d0d2 100644 --- a/tests/unit/llms/chatgpt/responses/test_chatgpt_responses_transformation.py +++ b/tests/unit/llms/chatgpt/responses/test_chatgpt_responses_transformation.py @@ -239,6 +239,7 @@ class TestChatGPTResponsesAPITransformation: "max_output_tokens": 123, "stream_options": {"include_usage": True}, # supported and should be preserved + "prompt_cache_key": "session_123", "truncation": "auto", "previous_response_id": "resp_123", "reasoning": {"effort": "medium"}, @@ -257,6 +258,7 @@ class TestChatGPTResponsesAPITransformation: assert "max_output_tokens" not in request assert "stream_options" not in request + assert request["prompt_cache_key"] == "session_123" assert request["truncation"] == "auto" assert request["previous_response_id"] == "resp_123" assert request["reasoning"] == {"effort": "medium"}