Release 0.6.6

Released 2019-02-11.

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

Release 69667 · Release 69668 · Release 31960

Back to the changelog