Code

History

Commits on v6.8.55 to test35+
  1. feat(thinking): improve provider family checks and clamp unsupported levelshkfires committed 8 months agoc80ab8b
  2. feat(thinking): enhance adaptive thinking support across models and update test caseshkfires committed 8 months agod2e5857
  3. feat(translator): support Claude thinking type adaptivehkfires committed 9 months ago938a799
  4. fix(translator): correct gemini-cli log prefixhkfires committed 9 months agob7e4f00
  5. Add Kimi (Moonshot AI) provider supporttest committed 9 months agof5f26f0
  6. fix(test): rename test function to reflect behavior change for builtin toolshkfires committed 9 months ago075e3ab
  7. feat(thinking): enable thinking toggle for qwen3 and deepseek models Fix #1245hkfires committed 9 months agoc8c2732
  8. feat(registry): support provider-specific model info lookuphkfires committed 9 months agoe641fde
  9. fix(executor): stop rewriting thinkingLevel for geminihkfires committed 9 months ago1d2fe55
  10. fix(thinking): update ValidateConfig to include fromSuffix parameter and adjust budget validation logichkfires committed 9 months agocb6caf3
  11. refactor(thinking): add Gemini family provider grouping for strict validationhkfires committed 9 months ago03005b5
  12. test(thinking): split E2E coverage into suffix and body parameter test functionshkfires committed 9 months ago97b67e0
  13. refactor(thinking): extract antigravity logic into a dedicated providerhkfires committed 10 months ago4ad6189
  14. feat(thinking): support zero as a valid thinking budget for capable modelshkfires committed 10 months agoff4ff6b
  15. test(thinking): remove legacy unit and integration testshkfires committed 10 months ago33d6695
  16. refactor: improve thinking logichkfires committed 10 months ago0b06d63
  17. Merge pull request #553 from XInTheDark/fix/builtin-tools-web-searchLuis Pater committed 10 months ago3d01b3c
  18. fix(thinking): fallback to upstream model for thinking support when alias not in registryhkfires committed 10 months ago8bf3305
  19. feat(thinking): add numeric budget to thinkingLevel conversion fallbackhkfires committed 10 months agod00e3ea
  20. feat(amp): add per-client upstream API key mapping supporthkfires committed 10 months ago225e2c6
  21. fix: apply thinkingLevel from model suffix metadata for Gemini 3Ben Vargas committed 11 months ago598f0af
  22. Revert "Fix invalid thinking signature when proxying Claude via Antigravity"Luis Pater committed 11 months ago7481c0e
  23. Merge pull request #570 from fuguiKz/fix/antigravity-thinking-signatureLuis Pater committed 11 months agof49e887
  24. Fix antigravity Claude thinking signature handlingkz committed 11 months agob602eae
  25. test(thinking): add effort to budget coveragehkfires committed 11 months ago9df96a4
  26. fix(thinking): align budget effort mapping across translatorshkfires committed 11 months ago28a428a
  27. fix(translator): preserve built-in tools across openai<->responsesMuzhen Gaming committed 11 months ago0b834fc
  28. refactor(thinking): export thinking helpershkfires committed 11 months ago367a05b
  29. fix(thinking): normalize effort mappinghkfires committed 11 months agod20b71d
  30. test(thinking): expand conversion edge case coveragehkfires committed 11 months agoa4a3274
  31. fix(thinking): centralize reasoning_effort mappinghkfires committed 11 months ago716aa71
  32. test(thinking): cover openai-compat reasoning passthroughhkfires committed 11 months ago8496cc2
  33. fix(thinking): gate reasoning effort by model supporthkfires committed 11 months ago5ef2d59
  34. fix(thinking): map budgets to effort levelshkfires committed 11 months ago374faa2
  35. refactor(api): simplify request body parsing in ampcode handlershkfires committed 11 months ago05cfa16