From 1908c3a6892d6d807ed177c88d3bed14bf9f74be Mon Sep 17 00:00:00 2001 From: Ishaan Jaffer Date: Wed, 8 Apr 2026 13:50:50 -0700 Subject: [PATCH] fix: context_management --- .../handler.py | 24 +++++++++++++++++++ 1 file changed, 24 insertions(+) diff --git a/litellm/responses/litellm_completion_transformation/handler.py b/litellm/responses/litellm_completion_transformation/handler.py index 5faa8b587c9..686481a4ac8 100644 --- a/litellm/responses/litellm_completion_transformation/handler.py +++ b/litellm/responses/litellm_completion_transformation/handler.py @@ -48,6 +48,8 @@ class LiteLLMCompletionTransformationHandler: **kwargs, ) + kwargs.pop("context_management", None) + if _is_async: return self.async_response_api_handler( litellm_completion_request=litellm_completion_request, @@ -105,6 +107,28 @@ class LiteLLMCompletionTransformationHandler: litellm_completion_request=litellm_completion_request, ) + kwargs.pop("context_management", None) + + context_management = responses_api_request.get("context_management") + if context_management: + compaction_model = litellm_completion_request.get("model", "") + compaction_provider = litellm_completion_request.get( + "custom_llm_provider" + ) + if compaction_provider: + compaction_model = f"{compaction_provider}/{compaction_model}" + request_input = await LiteLLMCompletionResponsesConfig._transform_context_management( + model=compaction_model, + input=request_input, + context_management=context_management, + ) + litellm_completion_request["messages"] = ( + LiteLLMCompletionResponsesConfig.transform_responses_api_input_to_messages( + input=request_input, + responses_api_request=responses_api_request, + ) + ) + acompletion_args = {} acompletion_args.update(kwargs) acompletion_args.update(litellm_completion_request)