# Limits are Ridiculous and needs to Change **URL:** \ **Category:** ChatGPT\ **Tags:** gpt-4, chatgpt\ **Created:** [June 3, 2024, 8:14pm UTC](https://community.openai.com/t/limits-are-ridiculous-and-needs-to-change/794324 "2024-06-03T20:14:12Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![bsl64](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/bsl64/32/451235_2.png) [@bsl64](https://community.openai.com/u/bsl64)\ **Post date:** [June 3, 2024, 8:14pm UTC](https://community.openai.com/t/limits-are-ridiculous-and-needs-to-change/794324/1 "2024-06-03T20:14:12Z")
Can’t even find out how to contact support. The chat keeps sending me in an endless loop and never lets me send a message. I like chatGPT 4 image generator much better than the 4o. Not sure why there is a different but there is. But I have been generating about 5 images and then getting a message saying I have reached the limit for gpt 4 and have to wait for 2 hours!!! What is going on? Surely I can create more than 5 images??? I have not used chatGPT for many hours, I just came back, created a few images, and got another rate limit for chatgpt 4.0! Why? This is crazy. Never paid for a service where I have to pay not to get to use it. ---
**Author:** ![anon22939549](https://avatars.discourse-cdn.com/v4/letter/a/f14d63/32.png) [@anon22939549](https://community.openai.com/u/anon22939549)\ **Post date:** [June 3, 2024, 10:36pm UTC](https://community.openai.com/t/limits-are-ridiculous-and-needs-to-change/794324/2 "2024-06-03T22:36:18Z")
> [@bsl64](#): > > I like chatGPT 4 image generator much better than the 4o. Not sure why there is a different but there is. The image generator is the same—Dall-E 3—the only difference is the model writing the prompt for you. > [@bsl64](#): > > But I have been generating about 5 images and then getting a message saying I have reached the limit for gpt 4 and have to wait for 2 hours!!! What is going on? Surely I can create more than 5 images??? `dall-e-3` is a (relatively) expensive model. Through the API a single image is $0.04–$0.08, or about what 1,300–2,700 `gpt-4-turbo` output tokens costs. That’s on the order of 3–9 typical message outputs. So, if you’re using the `gpt-4-turbo` model to call `dall-e-3` each message exchange uses 4–10 times as much compute. If there’s not a particularly heavy load at the moment so you getting the full 40-messages/3-hours I think you should be reasonably expecting to be able to generate roughly 4–10 images in that window if you’re not having any other pure text exchanges. I expect this math will change quite a bit once the native image-generating capability of `gpt-4o` is released. ---
**Author:** ![bsl64](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/bsl64/32/451235_2.png) [@bsl64](https://community.openai.com/u/bsl64)\ **Post date:** [June 4, 2024, 9:42am UTC](https://community.openai.com/t/limits-are-ridiculous-and-needs-to-change/794324/3 "2024-06-04T09:42:58Z")
I don’t know all you’re saying, all I can say is that I was able to generate images and use chatGPT 4 without these interruptions in the past. I am not using it any more than I did months ago and now, I am getting these messages and even got one today that said I had reached my daily limit for image generation! It said I had to wait 9 hours 13 minutes to use it again! Now that makes no sense as I never received that message before and I am not using this any more than I have in the past. It’s getting harder to use this all of the time. Makes no sense to pay for a service where your work flow is constantly interrupted. ---
**Author:** ![josephgossel](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/josephgossel/32/147383_2.png) [@josephgossel](https://community.openai.com/u/josephgossel)\ **Post date:** [June 4, 2024, 12:49pm UTC](https://community.openai.com/t/limits-are-ridiculous-and-needs-to-change/794324/4 "2024-06-04T12:49:02Z")
The interruptions are minimal compared to how hastily it mashes innane details into images now… after uploading 10 images of myself, it often parts my hair or adds a piercing of some sort. This model seems almost predatory, intentionally generating incorrect images or providing terrible data.