From 6f2fc2f1cbfbc0714e0349cb48ea19698f3f71c5 Mon Sep 17 00:00:00 2001 From: Stephen Hu Date: Sun, 4 Jan 2026 10:31:56 +0800 Subject: [PATCH] refactor:re order logics in clean_gen_conf (#12391) ### What problem does this PR solve? re order logics in clean_gen_conf #12388 ### Type of change - [x] Refactoring --- rag/llm/chat_model.py | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/rag/llm/chat_model.py b/rag/llm/chat_model.py index e1451c00d..a9e3c1ab7 100644 --- a/rag/llm/chat_model.py +++ b/rag/llm/chat_model.py @@ -100,6 +100,12 @@ class Base(ABC): return LLMErrorCode.ERROR_GENERIC def _clean_conf(self, gen_conf): + model_name_lower = (self.model_name or "").lower() + # gpt-5 and gpt-5.1 endpoints have inconsistent parameter support, clear custom generation params to prevent unexpected issues + if "gpt-5" in model_name_lower: + gen_conf = {} + return gen_conf + if "max_tokens" in gen_conf: del gen_conf["max_tokens"] @@ -127,12 +133,6 @@ class Base(ABC): } gen_conf = {k: v for k, v in gen_conf.items() if k in allowed_conf} - - model_name_lower = (self.model_name or "").lower() - # gpt-5 and gpt-5.1 endpoints have inconsistent parameter support, clear custom generation params to prevent unexpected issues - if "gpt-5" in model_name_lower: - gen_conf = {} - return gen_conf async def _async_chat_streamly(self, history, gen_conf, **kwargs):