xianshujun cee7ed4b32 fix(openai): keep answers in content when thinking is disabled ...

qwen3's chat template wraps <think> in {% if enable_thinking %}, but the
prefill detectors scan the raw template without evaluating the condition.
With streaming and no tools, the synthetic prefill made the Go-side extractor
classify the whole answer as reasoning whenever the client sent
metadata.enable_thinking=false.

Skip prefill detection when the request disables thinking. Real
<think>...</think> blocks that a model emits anyway are still extracted from
the content itself.

Fixes #12939

Assisted-by: ZCode:deepseek-flash
Signed-off-by: xianshujun <xianshupang@gmail.com>

2026-11-02 12:38:30 +09:00

..

addons

feat(addons): log the reason behind a generic denial (#12875)

2026-10-24 01:30:41 +02:00

anthropic

feat(addons): installable add-ons foundation and the Tables add-on (#12702)

2026-10-22 17:57:19 +02:00

elevenlabs

fix(sound): enumerate expansion keys (#12787)

2026-10-20 13:14:49 +02:00

explorer

feat: add distributed mode (#9124)

2026-03-30 00:47:27 +02:00

jina

feat(whisper): honor client cancellation via ggml abort_callback (#9710)

2026-05-08 01:44:47 +02:00

localai

fix(proxy): propagate model env and accept direct keys (#12940)

2026-10-27 10:27:45 +01:00

mcp

feat(addons): run the add-on owner as its own workload (#12867)

2026-10-24 14:29:39 +02:00

ollama

fix(ollama): report on-disk size for /api/tags and /api/ps (#11989)

2026-09-28 00:11:41 +02:00

openai

fix(openai): keep answers in content when thinking is disabled

2026-11-02 12:38:30 +09:00

openresponses

fix(responses): preserve replayed output_text

2026-10-25 08:43:05 +00:00