I use gpt 4 turbo for image recognition for a site I run. And when I have really heavy usage of the API but still well under my limits all my requests start timing out.
Is this a hidden throttling mechanism they have? I’d like to know so I could adjust or plan usage accordingly.
Welcome to the dev forum
Yes there are rate limits and they are transparent about it.
I know and I’m on tier 5 and am well below my daily limits. Are there hidden rate limits where heavy usage will cause throttling?
When I have really heavy usage my requests start all timing out. +30s for gpt-4-turbo to analyse an image.