Release 0.24.10

Released 2016-09-14.

• Added per-model rate limits
• Fixed token counting for long prompts
• Added support for custom upstream timeouts

Release 81887 · Release 81888 · Release 75812

Back to the changelog