Release 0.4.0

Released 2019-06-11.

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

Release 36637 · Release 36638 · Release 82187

Back to the changelog