I started using the images generated by the GPT-4 model on April 25th, and although the images are created with the 4.0 model, they are extremely blurry and of very low quality. Has anyone else experienced this issue?
You marked this as API, but I’m wondering if you meant ChatGPT?
If you’re using the API, are you setting the quality?
It may be trying to emulate focal distance in your image.
Here’s one I made yesterday after watching some guy on youtube fail to get ChatGPT to generate a glass of wine filled to the top. (2 prompts by the way, he clearly wasn’t trying)
Note how crisp some of those lines are (in particular the reflection of the light). I’d recommend trying to add something in your prompt to get ChatGPT to ensure the whole image is in focus.
