Heo, Sung 5d67e76788 fix(api): surface Ollama `reasoning` field for Qwen3 thinking models ... Ollama's Qwen3 thinking models stream the chain-of-thought in a `reasoning` field with an empty `content`. The OpenAI-compatible provider only recognised `reasoning_content`/`thinking`, so reasoning-only responses surfaced no text at all. Add `reasoning` to both the streaming (`ChunkDelta`) and non-streaming (`ChatMessage`) shapes and fall back to it when extracting thinking, preserving the existing thinking/text content-block separation. Covered by two new tests for the streaming and non-streaming paths. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>		2026-06-06 08:09:18 +09:00
..
benches	omx(team): auto-checkpoint worker-1 [1]	2026-05-15 10:30:06 +09:00
src	fix(api): surface Ollama `reasoning` field for Qwen3 thinking models	2026-06-06 08:09:18 +09:00
tests	fix: validate env model selection	2026-06-04 00:30:13 +09:00
Cargo.toml	US-011: Performance optimization for API request serialization	2026-04-16 11:11:45 +00:00