I am a pay-as-you-go user and I have waited 48 hours, which means I should get 3,500 RPM and 90,000 TPM. I almost never get over 60 RPM which is far from the limit. Also the TPM is fairly low too, so it can't be that. I am still hitting the error 429 - Rate limit reached for requests sometimes. It is almost like I still have the first 48 hours rate limit, but I have waited about 4 days now and I am still hitting the error regularly. Is this a bug or can I fix it somehow?
#Getting Rate Limited Even Though It Is Far From The Limit
8 messages · Page 1 of 1 (latest)
This doesn't necessarily mean you're hitting your rate limit. It also sometimes occurs under extreme load.
Thanks for the reply. So it does not have anything to do with my usage, but rather OpenAI's servers are being overloaded? Does it also means that I can keep scaling my project with more users (more requests) without more 429 errors?
It's possible. I can't say for sure. I'm not sure how rate limits are calculated over time.
All I'm saying is that 429 occurs erroneously on occasion. :P
Thanks, I was seeing the very same issue. Are we still getting charged though?
If you receive a 429 error, you should not be charged, no.
same problem here!!