Release 0.35.8

Released 2015-04-13.

• Added per-model rate limits
• Fixed token counting for long prompts
• Improved streaming reliability

Release 73017 · Release 73018 · Release 64878

Back to the changelog