Note: Muse Image is the agentic image model from Meta Superintelligence Labs. Learn more on the AI at Meta blog and the Muse Image model page.
We've added Muse Image for both generation and editing. Create images from a prompt, or compose and refine existing images with natural language—strong instruction following, accurate text, and volume-friendly pricing for ad variants, catalog shots, and personalization.
About Muse Image
Muse Image plans and self-checks before it returns a result. It follows detailed instructions faithfully and renders typography accurately.
- Text-to-Image: Generate from a prompt across nine aspect ratios (21:9 through 9:21). Up to 10 images per run.
- Image Editing: Edit and compose from 1–10 reference images. Describe the change—the model applies it and leaves the rest alone.
- Ad-ready volume: $0.013 per image, so variants, catalog shots, and personalization stay affordable at scale.
What's New
- Muse Image in the image model selector, as a static-ad option for standard aspect ratios, and as an edit model wherever image editing is supported.
How It Works
Generation: Select Muse Image, enter your prompt, and choose an aspect ratio. Best when you need precise instructions and readable text on the image.
Editing: In the image editor, select Muse Image, add one or more source images (up to 10), and describe the change—composition, object edits, or refinements.