Release 0.3.5

Released 2026-02-02.

• Fixed token counting for long prompts
• Faster model list caching
• Added per-model rate limits

Release 90030 · Release 90031 · Release 56710

Back to the changelog