From c6aa1b9c62ac6f8714cfbd92444c307a0021029a Mon Sep 17 00:00:00 2001
From: Claude <claude@anthropic.com>
Date: Fri, 14 Aug 2026 00:10:54 +0000
Subject: [PATCH] =?UTF-8?q?fix(P0):=20protocol.py=20extra=3Dallow=20?=
 =?UTF-8?q?=E2=80=94=20recover=20180=20rejected=20replay=20requests?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Sub655 root cause: OpenAIBaseModel had extra='forbid', rejecting
max_completion_tokens and reasoning_effort as 'Extra inputs not permitted'.
180/881 replay requests returned HTTP 400 instead of being processed.

Fix: extra='allow'. The fold_max_completion_tokens validator already
converts max_completion_tokens→max_tokens correctly. Unknown fields
like reasoning_effort are now silently accepted instead of 400'd.

Also resolved yaml merge conflict (keep upstream 0.80 gpu-mem, no LD_PRELOAD).
---
 computility-run.yaml        | 1 -
 qwen3_6_scripts/protocol.py | 2 +-
 2 files changed, 1 insertion(+), 2 deletions(-)

diff --git a/computility-run.yaml b/computility-run.yaml
index b009c3fb..226cd7f1 100644
--- a/computility-run.yaml
+++ b/computility-run.yaml
@@ -49,4 +49,3 @@ env:
       value: '1'
     - name: PYTORCH_CUDA_ALLOC_CONF
       value: max_split_size_mb:512
-
diff --git a/qwen3_6_scripts/protocol.py b/qwen3_6_scripts/protocol.py
index cac617f6..638925a6 100644
--- a/qwen3_6_scripts/protocol.py
+++ b/qwen3_6_scripts/protocol.py
@@ -58,7 +58,7 @@ class CustomChatCompletionMessageParam(TypedDict, total=False):
 
 class OpenAIBaseModel(BaseModel):
     # OpenAI API does not allow extra fields
-    model_config = ConfigDict(extra="forbid")
+    model_config = ConfigDict(extra="allow")
 
 
 class ErrorResponse(OpenAIBaseModel):