kuku 845af59301

fix(models): preserve MiniMax reasoning_content responses (#6071) ...

* fix(models): preserve MiniMax reasoning_content responses

* fix(models): preserve whitespace MiniMax reasoning deltas

* test(models): pin MiniMax reasoning precedence

---------

Co-authored-by: Willem Jiang <willem.jiang@gmail.com>

2026-10-09 16:00:32 +08:00

..

__init__.py

refactor: split backend into harness (deerflow.*) and app (app.*) (#1131)

2026-03-14 22:55:52 +08:00

AGENTS.md

fix(vllm): normalize thinking switches without mutating request bodies (#6496)

2026-10-08 20:41:40 +08:00

assistant_payload_replay.py

fix(runtime): harden model response recovery at provider boundaries (#5080)

2026-09-18 07:29:34 +08:00

claude_provider.py

fix(middleware): guard OverflowError when parsing non-finite Retry-After hints (#6455)

2026-10-08 18:16:21 +08:00

credential_loader.py

fix(models): skip undecodable Claude descriptor handoffs (#6323)

2026-10-05 20:21:31 +08:00

factory.py

fix(models): deep-merge when_thinking_enabled on the legacy path (#5894)

2026-09-27 20:52:13 +08:00

mindie_provider.py

fix(models): align MindIE synchronous streaming compatibility (#6497)

2026-10-08 19:30:44 +08:00

openai_codex_provider.py

fix(models): preserve explicit Codex function tool strictness (#6360)

2026-10-06 16:53:45 +08:00

patched_deepseek.py

fix(runtime): harden model response recovery at provider boundaries (#5080)

2026-09-18 07:29:34 +08:00

patched_mimo.py

refactor(provider): share assistant payload replay matching (#3307)

2026-05-29 23:05:59 +08:00

patched_minimax.py

fix(models): preserve MiniMax reasoning_content responses (#6071)

2026-10-09 16:00:32 +08:00

patched_openai.py

fix(config): correct Gemini official endpoint reasoning example (#6161)

2026-10-02 22:32:21 +08:00

patched_stepfun.py

feat(models): add StepFun reasoning model adapter (#3461)

2026-06-09 18:01:43 +08:00

reasoning.py

feat(models): add declarative reasoning capability contract (#5780)

2026-09-24 20:39:30 +08:00

request_admission.py

fix(models): preserve FIFO across request-admission handoff (#5459)

2026-09-16 22:07:42 +08:00

vllm_provider.py

fix(vllm): normalize thinking switches without mutating request bodies (#6496)

2026-10-08 20:41:40 +08:00