I’ve been using the OpenAI PLUS service and accessing the ChatGPT-3.5 and GPT-4 models via an API key for about a year. Starting from yesterday, my self-hosted site can no longer access the API service. Every query I send returns the message: “You exceeded your current quota, please check your plan and billing details.”
When I checked the account management page, I noticed that my usage limit shows as zero, which seems to be the reason why I’m getting the “exceeded your current quota” message for any query.
However, my actual settings should not be zero. I tried resetting the limit, and although it says “Your limits have been updated,” the change doesn’t reflect on the Usage page.
Could someone help me understand what might be going wrong here?
I will add that I have observed the issue with the limit display on the usage page, too, when explicitly looking at the usage page for the default project. However, as of now the APIs still work for me without returning an error.
Under the usage page at the organizational level, my limits are displayed properly and I assume that the limits I have set at the organizational level still apply under the default project despite it not displaying properly.
That said, I encountered a limit related error yesterday when trying to submit a batch request and I now wonder that there might be a link between the error and this issue.
I have tried finding a solution but so far to no avail.
Same here. Haven’t received a rate limit warning yet but I’m thinking this is connected to the Projects setting that was launched recently. We are using the Default Project which states: “You cannot set limits on a default project. Create a new project to set limits.” so I’m assuming that we are allowed 100% usage on this project.
Thanks for the discussion. Honestly, I had just typed out a long message detailing my issue, but before sending it, I decided to check my site again and found it was back to normal.
I haven’t done anything since my post yesterday… just created a new project earlier because it said that limits couldn’t be set on the default project, but that didn’t solve the issue at the time.
Maybe it was just a silly bug, or perhaps creating a new project took some time to take effect, whatever.
Thanks for sharing that. Yes, I also created a new project yesterday, but it didn’t solve the issue at the time. However, just now I tried again and the API service is back to normal.
I’m not sure if it’s because of the new project or if OpenAI fixed some bugs introduced by an update. Anyway, it doesn’t matter now.
Well, then you discover in projects, there is no limiting billings anyway unless you are a mysterious “enterprise” API partner. The little guy has to pay big.
Glad to see they straightened out the API temporarly made unusable for some.
