Release 0.26.6

Released 2016-06-26.

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

Release 3907 · Release 3908 · Release 17394

Back to the changelog