Skip to content

feat(provider): align reasoning defaults and add kimi-k2.6 to catalog#160

Merged
duguwanglong merged 2 commits intomainfrom
feat/provider-reasoning-defaults-kimi-k26
Apr 22, 2026
Merged

feat(provider): align reasoning defaults and add kimi-k2.6 to catalog#160
duguwanglong merged 2 commits intomainfrom
feat/provider-reasoning-defaults-kimi-k26

Conversation

@xiami762
Copy link
Copy Markdown
Contributor

@xiami762 xiami762 commented Apr 22, 2026

Summary

  • Respect enable_thinking from model settings when building provider options (Claude/OpenAI/Gemini/Groq/Bedrock/Qwen/Kimi paths).
  • Parse optional max_input_tokens from the curated catalog; add kimi-k2.6 entry for threatbook-cn-llm.
  • WebUI Model sheet: align the reasoning toggle with server defaults and persist enable_thinking in default_parameters on save.
  • Tests: refresh curated catalog expectations (e.g. Anthropic model ids, cherry provider) and add test_provider_options.py.

Honor enable_thinking from model settings when building provider options;
parse max_input_tokens from catalog; add Moonshot/Kimi K2.x handling.
Update WebUI model sheet to match defaults and persist enable_thinking.
Sync curated catalog tests and add provider options coverage.

Made-with: Cursor
@xiami762 xiami762 requested a review from duguwanglong April 22, 2026 05:55
Keep curated provider coverage aligned for kimi-k2.6 and add bootstrap tests so the model is available during provider setup.

Made-with: Cursor
@duguwanglong duguwanglong merged commit 21490a3 into main Apr 22, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants