Yesterday I requested higher monthly usage limits as we are launching a SaaS and can’t have spend limits.
Today in our account I can see that in our limits:
“You must be on a paid plan with arrears billing to manage usage limits.”
Do you think it’s their way of removing us from a paid plan to give us limitless spending? I had no response from them, just this poped out in our account.
As you spend more through the API and use it more your limits will raise, there are millions of developers, lots of whom make requests for large limit increases for their products, OpenAI have to balance usage against the available hardware to run those requests on.
The best way to be able to apply for increases is to have a proven track record of using up to your limits, but remember this is a limited resource.
The developer forum does not have access to users accounts, you can contact support on help.openai.com and use the contact bot in the bottom right corner and leave your contact details and a description of the issue.
shows a list of models and their per-minute limit by tier
Usage Limits
Manage your API spend by configuring monthly spend limits…
Has the maximum usage OpenAI allows for your organization each month, along with soft (email) and hard limits you set yourself
Increasing your limits
Has your tier and the qualifications for the next tier
It’s possible there’s a “valued partner, skip all that” monthly limit checkbox on OpenAI’s backend for if you a prepaid, trusted, and are filling your account with non-refundable credits.
A hard limit can prevent excessive usage. It can prevent costly account abuse in cases where an API key was hacked, or unexpected charges in general. If this was made the interface for all developers that have a prepaid API account, that would seem to mean that there was nothing preventing the entire account balance from being spent by such undesired usage - or AI agent run amok.
Our tier 5 organization is also experiencing the same issue. We cannot modify our soft or hard threshold limits. This is problematic as our maximum usage limit is 10k USD per month; however, we have a threshold set to 2k which we will exceed very quickly if we cannot change. Reaching out to OpenAI support has been unhelpful as I am receiving canned responses from their bot basically stating to proceed with a pay as you go plan. We are on a billing plan for many months at this point. Any idea?
I am also receiving this message. I think it may be a bug, since I don’t understand why we’d be getting this message otherwise. We are also a tier 5 organization. Can OpenAI support staff assist us please? This is an urgent request, since we are aiming for an upcoming conference deadline for our research but will likely hit our API limit so we need to increase our monthly quota…
I routinely crash my monthly limit down to my current spending level when I don’t expect to use the API for a while, then bring it back up when I’m going to use it again.
It is likely the option for a self-set limitation on spending is now permanently gone. What you previously had for email notification or cutoff for unexpected abuse-like usage is no longer in effect.
If you aren’t being cut off by an unseen and unalterable personal monthly usage limit, that part is not a concern. There was only one day of a user’s inaccessible settings blocking API calls reported on the forum.
The message that you can’t have personal limits didn’t get there except by obvious choice by OpenAI.
A user-set cutoff is no longer bounded and delineated by months and a maximum monthly bill which prepaid users never get. That part sort of makes sense. (although tier still enforces a monthly limit, still by month-end)
The effect is your account is a free-for-all for API key leaks and account hacks. And your experiments with “assistants” gone out of control.
The cutoff now? When your account credits are completely emptied. Or when “auto refill” has emptied your bank account.
(boardroom conversation: “How can we sell more cars??” “Remove the brakes?!”)
I am getting the error 429. - You exceeded your current quota. Interestingly, this error only shows up when I try to access the OpenAI API with a developer key. However, the interactive chat works just fine. I am not sure how and which business logic is handling this!
You can see the prices for AI language data use here: Pricing
ChatGPT is a consumer product on the web, and OpenAI offers the basic version for free, and ChatGPT Plus for $20/per month. ChatGPT is unconnected to the API billing.
You must purchase prepayment of credits for your organization account before you will be able to consume API services.