Algorithm of OpenAI's rate limit policy

Hello everyone,

I have a question regarding the rate limit algorithm used by the OpenAI API. I am curious to know if anyone has information or resources about the specific algorithm employed. I believe it could be one of the following:

  1. Fixed window counter
  2. Sliding window log
  3. Sliding window counter

After reviewing the public documentation, I couldn’t find any disclosure regarding this information. If anyone has insights or resources that can shed light on this topic, I would greatly appreciate your input.

Thank you very much!

Here you find a documentation about Rate limits | OpenAI Help Center with the API.

I hope this helps.