What's Changed
- fix: respect enableMaxTokens setting when maxTokens is not configured by @0xfullex in #11438
- fix: websearch button condition by @DeJeune in #11440
- ⬆️ chore: upgrade @anthropic-ai/claude-agent-sdk to 0.1.53 by @vaayne in #11444
- fix: correct updateAssistantPreset reducer to properly update preset by @0xfullex in #11453
- fix(models): qwen-mt-flash supports text delta by @EurFelux in #11448
- fix: Groq verbosity setting by @EurFelux in #11452
- fix: header merging logic via chore ai-sdk by @DeJeune in #11443
- Fix: custom parameters for Gemini models by @Copilot in #11456
- fix: update Azure provider handling in AI SDK integration by @DeJeune in #11465
- docs: enhance CLAUDE.md with quality guidelines by @EurFelux in #11464
- fix: improve Gemini reasoning and message handling by @DeJeune in #11439
- fix: add claude-opus-4-5 pattern to THINKING_TOKEN_MAP by @luosc in #11457
- feat: add silicon provider support for Anthropic API compatibility by @DeJeune in #11468
- fix(anthropic): prevent duplicate /v1 in API endpoints by @EurFelux in #11467
- fix: get quota and quota tips by @eeee0717 in #11472
- Fix Poe API reasoning parameters for GPT-5 and reasoning models by @Copilot in #11379
- fix: silicon provider code list by @DeJeune in #11474
- fix(i18n): clean up translation tags and untranslated strings by @EurFelux in #11471
- feat(aihubmix): support nano banana by @DDU1222 in #11476
New Contributors
Full Changelog: v1.7.0-rc.2...v1.7.0-rc.3