Image Generation & Editor

Generate an image from a prompt, mask a region and regenerate just that part, or ask the model to describe an image in words.

gpt-image-1-mini · text-to-image
Enter a prompt and press Generate.
Your generated image will appear here.
40
5 · fine120 · broad
Painted area will be regenerated
gpt-image-2 · masked edit (inpaint)
Load an image. Paint a region to change just that, or Regenerate to rebuild the whole image.
Load an image to edit
choose one below, or send one from Generate
·

Versions · click one to keep editing it

Load an image and regenerate — each version shows up here.
Load an image to describe
choose one below
·
gpt-4o-mini · image-to-text (vision)
Load an image, then press Describe.
The model's description will appear here.

Source on GitHub. Sample photo by Alvan Nee on Unsplash.