BUG: Rate limit broken? - Slow GPT3.5 and GPT4 - But GPT3 is fast!

I see many others seems to have speed issues with the API today.
I get rate limit warnings after maybe 500 tokens:

“Rate limit reached … Limit: 90000 / min. Current: 87859 / min.” I get this within seconds even if I have been idle.

3.5 and 4 seem to have the same issue. - But 3 is fast, as 3.5 was before. This for the same API key.

I am guessing this is just a temporary issue.

Yes, that would be my guess. You can always make a report to help.openai.com (support bot bottom right)

1 Like

It sounds like the same misdetection as a day ago – just not throwing errors that crash applications this time.

1 Like