From 9ba7253094d773a80e46edc10a4df19d07909013 Mon Sep 17 00:00:00 2001 From: gongwei-130 <56567052+gongwei-130@users.noreply.github.com> Date: Thu, 21 Aug 2025 13:22:03 -0700 Subject: [PATCH] accomendate reasoning_effort set in chat_template_kwargs (#9458) --- python/sglang/srt/entrypoints/openai/serving_chat.py | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/python/sglang/srt/entrypoints/openai/serving_chat.py b/python/sglang/srt/entrypoints/openai/serving_chat.py index d87c50dd6..042911f63 100644 --- a/python/sglang/srt/entrypoints/openai/serving_chat.py +++ b/python/sglang/srt/entrypoints/openai/serving_chat.py @@ -87,6 +87,14 @@ class OpenAIServingChat(OpenAIServingBase): self, request: ChatCompletionRequest, ) -> tuple[GenerateReqInput, ChatCompletionRequest]: + reasoning_effort = ( + request.chat_template_kwargs.pop("reasoning_effort", None) + if request.chat_template_kwargs + else None + ) + if reasoning_effort is not None: + request.reasoning_effort = reasoning_effort + """Convert OpenAI chat completion request to internal format""" is_multimodal = self.tokenizer_manager.model_config.is_multimodal