Finally reached the usage cap for GPT-4

See my post on subscriptions. Private usage, other subscription models

Yes, I also hit the cap whilst developing GPT’s. I was also experiencing the cap using plugins. It’s really annoying. If AI is ment to multiply your capabilities, then it’s a bummer to be limited by an subscription fee without any other options.

BTW: my employer is an Enterprise subscriber. I have professional access to ChatGPT. But I don’t want to proces my personal queries over my employer’s infrastructure. From GDPR perspective my employer is not allowed to do so. My profession is privacy officer.

I’m happy to pay twice the fee for twice the capacity. Even three times the fee to triple the capacity. What can I say? It’s a simple flat type of subscription model with different limits. OpenAI has already the functionality. It only some configuration I assume and a sales page for different private subscription models. How hard could the implementation be? Increased revenue for OpenAI because I have seen equivalent requests many times.

And you know what OpenAI? Because my private ChatGPT learnings I’m now asked to assist with the development of a specific GPT for corporate usage. Which will again result into increasing revenues for OpenAI. What’s the problem OpenAI?

I’ve read somewhere a response of OpenAI: “There aren’t enough GPU’s/TPU’s in the world to fulfil API developers’ needs.” I can understand that, but from my perspective it’s only me, myself and I. We are limited in our resources to make extensive use of ChatGPT. It’s at any time only a single user. If this user is not walking the dog. Our GPU/TPU usage is limited.

I also experienced the cap whilst developing a GPT but I also still could use ChatGPT 4. There’s clearly a discrepancy between both services regarding limits.

1 Like

Yeah - that is new.

I knew what the usage cap was before, even intuitively, but now… I cant even predict it.
But it is certainly far short of what it was before.

I cant tell if its a combination of 24 usage and 3 hour usage, or if its just overall… what?
I dont know.
But your not crazy.

1 Like

Thanks for this.

I was searching for at least an hour know to get to the root of the cap problem I experience. I suspected that there is a different cap for GPTs, as have other users in other threads on this forum.

However, no one provided actual proof of this.

What makes me wonder though, is the fact that I solely used GPTs today and I was also capped for usual GPT-4. I was definitely capped at around 20-25 messages, so this number seems accurate. According to their logic though, I should still be able to use GPT-4 for either 40 messages, or 25.

So this does not seem to be true.
I assume this was an official reply from the OpenAI staff?

To avoid wasting time when I hit my usage cap, I switch to https://platform.openai.com/playground, though each interaction costs something.

Yesterday was worse I barely counted about 20 messages.




Funny I only started getting these messages after I subscribed and what’s worse is that I don’t even use 40 prompts then I get this. So I don’t know if OpenAI includes responses in the 40 cap limit. I think it’s ludicrous that the limit should be set for paying subscribers.

1 Like

New Paid User here, just signed up eventually, never hit a cap before using free version (not that I remember) love using it.

Now I’ve been bricked from using it, even normal chat, which is confusing, because Microsoft Copilot is FREE on windows 11 for me and I can use that with no limits on image generation using Dalle-3 and chat to GPT 4 indefinitely.

I thought this is rather strange, I can’t continue my GPT at all now, just says it’s blocked, no warning, no timer or credits showing or anything on prompts, just a strict message popping up saying I can’t use it till tomorrow night which is hours away? I’m quite disappointed as I thought I could use this freely especially after subscribing to it :confused: - also it can’t read from a Google Drive public link or folder either it states?

Not really been using it much today but hit a hard limit when uploading some PDF’s training a GPT to look at invoices - what I was surprised at is that it cannot use a single zip file for me to upload to look at all the PDFs for extracting the data? also cannot read the PDF’s earlier said something went wrong, then later tonight could all of a sudden read them, I feel like now I wasted so many prompts earlier and attempts at trying to get it to accept the PDF’s due to a “system error” and now getting bricked. I even spent time merging all the PDF files into just one with all the data on there, now can’t even upload one file.

What I would suggest, especially for new user experience, is to at least have some sort of meter that we can visually see on the limitations and progress of how many prompts we have if there are restrictions, I did not even know there were any restrictions on the amount of prompts when initiating the sign up process and payment.

1 Like

I contacted support about this, new paid user, got locked out fairly quickly felt a but odd as I had just paid, was previously using windows 11 built in Chat GPT-4 and has very little limitations on mostly everything so really doesn’t make sense?

They replied that Microsoft is a different entity etc and that they use different servers? I thought Microsoft more or less owned Open Ai correct me if I’m wrong?

Will I lose all of the work I did up to the point when it crapped out? I think I reached the limit because I kept asking it to redo something and it kept redoing something else. I’m not sure how to approach the problem. I have two questions: how long before I’m back up?
Is there some language that chat 4 understands when I ask it to bring back something it did before? It kept bringing back the wrong thing.