Rate limit for V1 API ?

[yiwangethan]

May I ask what's the rate limit per account and per address of V1 API?

[acheong08]

Per account is a variable by OpenAI. It is estimated to be 50-70 requests per hour. Per IP address is 50req/minute

[linhandev]

Just tried 60 req/hour and 120 req/hour on a paid account. Both didn't hit rate limit.

[acheong08]

Paid account got much higher limits

[yiwangethan]

Can I just add a parameter "paid" in the developer API to declare that this is a paid account? Like below:
chatbot = Chatbot(config={
"email": "",
"password": "",
"paid": True
})

[linhandev]

Can I just add a parameter "paid" in the developer API to declare that this is a paid account? Like below:
chatbot = Chatbot(config={
"email": "",
"password": "",
"paid": True
})

There won't be any bug if you set paid=True on unpaid account, but the model used during inference will still be the unpaid model

[yiwangethan]

Thanks for your reply. I tested a paid account, and the rate limit for a paid account is about 300 requests/hour.

[hrizal]

hmm per IP Address ,,, noted

[acheong08]

That is on the proxy server's side to limit abuse