v4.5.0
Notes for v4.5.0
Changes
Multiple Rate Limits
The API now supports multiple rate limits to allow burst applications more overhead.
To support this, we’ve implemented the IETF Draft for RateLimit headers to provide a clear indication of limits within the headers.
As this is a draft, specifications may change, but it is currently the best way to display multiple limits.
This adds two new headers:
ratelimit-policyandratelimit.Example:
ratelimit: "per1min";r=86;t=58;pk=:NzExYTliNTMtYzc1ZS00ZjI2LTk2ZGMtNDg4NmI2OTc4NTlj:Example:
ratelimit-policy: "per1min";q=90;w=60;pk=:NzExYTliNTMtYzc1ZS00ZjI2LTk2ZGMtNDg4NmI2OTc4NTlj:
For more info, check the IETF Draft or ask an LLM using the documentation content.
The X-Ratelimit-* headers will continue to show the lowest/nearest rate limit to reach, for 99% of users will therefore nothing change
MMR v3 Peak RR
Now defaults to
0. Check this thread to find out why: Incorrect PEAK MMR
Last updated