Waiting Time of Rate Limit Error

API
1

Hi guys,

I have some question about Rate Limit Error.

The question is: How long I have to wait before I can use the OpenAI API again when triggering the rate limit?

After reading the documentation, I think there are two possible outcomes:

  1. Once triggered, the API will be locked for one minute.
  2. The slicing window is set to one minute, and the system continuously calculates the total number of requests and tokens used in the past minute. Therefore, even when the rate limit is triggered, the system will still calculate the past minute for each moment according to the slicing window. As a result, the unlock time may vary.

Thanks a lot for answering my questions!

2

hi
Check your rate limits - depending on the account for each user - in your platform rate limits: Rate limits - OpenAI API

image
image2239×667 68.7 KB

In this example 3 requests per minute (RPM). Other models have different rate limits.
I hope this helps.