Suggestion: Allow us to start using API key (to pay to continue) when chatGPT quota runs out

The chatGPT interface is extremely convenient. But it is very annoying when the 25 messages runs out. In such case, one needs to either switch to using a third party script/program to interact with the API, or the API playground.

It would be nice if we could choose to continue using chatGPT, but just use our paid API key as necessary.


I agree that there needs to be something, buuuuut. at the moment it’s down to resources, there is not enough compute on racks to serve all the extra API calls that would be made if such a service was offered.

The only thing I can say is 7.2mins. If you (on average) wait 7.2 mins between GPT-4 goes, you never run out.

1 Like

I think this is extremely unlikely to ever happen.

ChatGPT is a consumer product while the OpenAI API is an enterprise product.

It would be very unusual for a company to intermingle their products like that.

Yea, I think the real solution will be when ChatGPT for Business launches.

and what do you expect that to be? a web UI where you can pay for more than 25 messages per day and integrate other tools?

I imagine it will probably include a site license to cover multiple employees, allow easier sharing of chats between users including real-time collaboration with 2–5 users active in a single chat, and have some type of hybrid-pricing model which is more similar to API pricing than ChatGPT pricing (along the lines of 100M tokens for $1000/month) then pay-as-you-go after). It’ll also include enhanced support, access to more specialized models, and fine-tuned GPT-4.

1 Like

You think it will have similar rate / spend limits as the API?

No, if I was guessing I would expect the rate would be at a substantial discount to the API.

As it stands, a person could theoretically squeeze about $40/day worth of tokens out of GPT-4 using ChatGPT.

So, I’m assuming the rate for businesses will be at a much lower rate than the API, because if a business were to get one ChatGPT Plus with of use for one user 160-hours/month paying API rates, that would be on the order of $250–$300/month.

So, the rate will need to be much lower probably an order of magnitude lower, then with the rest of the monthly fee going to support costs.