How are rate limits aggregated?

I have a question about API rate limits. Simple question.
Are rate limits applied on a per API key basis?
Or is the limit applied to the entire organization in the aggregate?


1 Like

Based on the organisation.
The backoff strategy given in the OpenAI documentation is really helpful in trying to not encounter this problem if you are using it as a part of a pipeline.

1 Like

Thank you . I understand.
@novaphil @udm17