Does OpenAI adjusted CAP?

((You’ve reached the current usage cap for GPT-4, please try again after…)))

Last several interactions with GPT4 while building assistant I got in to cap. First times havn’t checked but after it felt too quickly I double checket our interactions and I have not used even 20 messages. While messages were quite complicated and demanding compute resources it still wasn’t 20 messages. Isn’t the cap 50 messages per 3 hours?

I’m I alone facing this weird cap? Or does OpenAI making tricks and calculates not only messages but compute usage and don’t actually provode on what they giving word ?

1 Like

I just got capped at 11 messages. You can’t even finish a full conversation.


Yes, they changed the message cap

I’m not a fan of that… With the shop’s update, it’s not enough
They do it because the infrastructure is expansive
I wish they had a higher personal tier for, like 40$ that will allow 100 messages per 1 hour or something like that

1 Like

I want to pay extra for unlimited usage.


That’s what the API is for, but you must pay per use and build your own interface.

I do not use API, but GPT-4. Today, after a long time, I encountered a limit again when using ChatGPT4.

1 Like

Is the cap only for free tokens that openai gives or its also for the tokens you buy?

1 Like

This three-hour usage cap discussed in the topic, which can be dynamic at times of high load, refers to only ChatGPT Plus while using GPT-4 within. ChatGPT Plus is the consumer chatbot on a webpage at which can be upgraded for a monthly subscription.

The API is different, for programmers, and has some rate limits, but the usage is basically “how much do you want to spend” - because you are billed for the language data of each API call. Access to GPT-4 models is unavailable while only using free trial credits.

1 Like

okay I understand now, that’s a relief. thank you so much

1 Like

This seems extremely shady. Why don’t you show how many prompts a user has done and indicate how close they are to hitting this limit, or warn them before they get close? It feels likes its more an arbitrary number that you decide. It doesn’t help that we are paying for this service. Does anyone know how this compares to gpt-3.5 prompt limits? If this has a higher prompt limit which I suspect it does, it would be great if someone could help me understand whats going on here

i guess such infos are important to customers before they pay, at least in europe are regulation about such things…

and my very first 5 minutes on chat 4 ended with a cap after just a few requests / questions
( who are also in the worst way possible answered )

instant cancelation

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.