Finally reached the usage cap for GPT-4

Anybody else reached the limits?

It happens only when I access GPTs. Never happened to me before. That may indicate how hot this GPT thing really is.

If there are enough people reporting the same problem, the OpenAI DevOps guy may notice it to plan better for the debut of the GPT Store at the end of the month. Just imagine the traffic. Another technology revolution may be just unfolding…

Denied

12 Likes

Every day happend to me.
This is the beauty :grinning: :grinning: :grinning: :grinning: :grinning:

6 Likes

Yup, I believe that’s just the old 50 prompts/3 hours limit that has been in place for months now.

It’s annoying of course, but that just forces you to learn to prompt well lol.

1 Like

On a related note, after setting up a GPT via my Plus account, I received email from OpenAI saying “You’ve used up all the credits available in your free trial.”

I didn’t think I submitted 50 messages, but does this imply that GPTs are subject to API usage limits even if using them within the ChatGPT “Plus” subscription account?

Hi and welcome to the Developer Forum!

The two events are purely coincidental, GPTs are covered under your ChatGPT Plus membership.

1 Like

Thank you. Since I am not currently using platform directly, is there another reason why I would have received that email saying that I’ve used up all my free credits? Is this just a marketing email from OpenAI trying to get me to sign up for their platform API service?

My guess is that it has been 3 months to the day since your opened your OpenAI account, free credits last for 3 months.

Thanks. One additional related question, why do 2 organizations show up in my platform account under my name/email address. One org shows up as the company incorporated name and the other org is the dba name.

They both show up under “Organization” and have separate IDs. Is there any reason why 2 orgs should or need to show up here, versus just having one org name and ID?

new subscriber for 2 days then had this message for 3 times. Every time need wait 3-6 hours :woozy_face:

How to solve this problem?
:sweat_smile:

Yes me too, its a bit unfortunate. I wouldn’t mind paying the double per month but having a full access without denial of access due to traffic or usage cap.

3 Likes

Maybe just create two accounts, it will cost you twice and much and try to not use GPTs in your main account. I think it may be causing these issues

1 Like

Not sure why that may be happening, at the moment, the account support site is down, when it is back up you should ask regarding that on help.openai.com (bottom right corner icon)

yes same here. It will be actually helpful to have some sort of indication - like a timer to know when I am reaching the threshold. I was just working on something for tomorrow and this came out of blue. I could not pace for it because I never reached cap before. Probably waste a lots of unecessary prompts. now I know.

When GPT-4 first introduced, it had an prompt indication on the bottom of page which was quite helpful as I would organize my prompts and pace accordingly. if it is true that there is a 50 prompt limit per 3 hours, then a prompt limit indicator or counter like the earlier day might be helpful. currently, there is no indication of any sorts

1 Like

Yes, I’ve also just run into it while setting up a GPT.
@Foxabilo Do you know if updating the GPT instructions eats up interactions, or is it just that when iterating on a GPT’s instructions it’s common to test each iteration which quickly gets you to the 50 message cap?

Changing settings will require making a call to the API, I’ve not tested if setting updates are counted against your rate limit, my guess is that everything is counted as the system is there to stop abuse and setting updates could be abused.

I suggest you get 2 accounts. Then you can log in and log out appropriately

That won’t help. Our GPT is to do design and development work. The state is associated with each OAuth authenticated user.

Well, technically, they could exchange information. But, it would be so lame to have an architecture like that just to work around the GPT capacity limits.

1 Like

I was thinking more in regard to my usage as my wife has an account because currently I am mainly creating videos on how to use my custom GPT tools. Yet I did realise that I couldn’t develop new ones or tweak them in that scenario