Meta · AI model
Muse Image
Meta Superintelligence Labs' image generation and editing model. It reasons before rendering and uses built-in web search, image search and code execution. On the Meta Model API it costs a flat $0.01 per generated image; it is also available in Meta AI.
Last verified Preview 4 sources
Key specs
- Provider
- Meta
- Family
- Muse Image
- API identifier
- muse-image-1.0
- Availability
- Preview
- Release date
- 2026-07-07
- License
- Proprietary
- API access
- Yes
- Consumer app
- Yes
Capabilities
Inputs
- Text input Yes
- Image input Yes For editing: 1 to 10 input images on the images edits endpoint
- Audio input No Inputs listed as text and image only
- Video input No Inputs listed as text and image only
- File input Unknown
Outputs
- Text output Unknown
- Image output Yes
- Audio output Unknown
Features
- Reasoning Native Reasons before it renders. reasoning_strength 'high' (default) refines over several passes; 'low' returns after one pass. No off setting is documented
- Tool calling No Image models only support the image_generation tool; function tools return HTTP 400
- Web search Yes Built-in image and web search grounding, on by default and switchable with enable_image_search / enable_web_search; the web_search tool itself is not accepted
"Unknown" means we have not verified this capability yet; it does not mean the model lacks it.
Pricing
Estimate your monthly cost →| Dimension | Tier | Price | Since | Verified | Source |
|---|---|---|---|---|---|
| Image generation | base | $0.01 per image | 2026-10-12 | 2026-10-25 | Pricing and rate limits - Meta Model API Flat price per successfully generated image, regardless of prompt length, reasoning strength or tools used; failed or safety-filtered images are not billed. Default size 'auto' (default aspect ratio); no quality options are documented |
Prices are shown in the currency the provider publishes; we never convert currencies.
Compare Muse Image
Guides
About Muse Image
Muse Image is Meta's image model, announced on 7 July 2026 and available in Meta AI and on the Meta Model API (muse-image-1.0). One model handles both generation from text and editing of existing images, including multi-turn editing on the Responses API and composition from multiple reference images. It reasons before rendering, and its built-in web search, image search and code execution are on by default. The API charges a flat $0.01 per successfully generated image.
Strengths and weaknesses
Strengths
- Flat $0.01 per generated image, regardless of prompt length, reasoning_strength or tools used
- One model for generation and editing; 1 to 10 input images per edit and up to 10 images per request on the images endpoints
- Built-in web and image search grounding and code execution, with no separate search charge
- Output as WebP (default), PNG or JPEG
Weaknesses
- The size parameter sets the aspect ratio, not exact pixel dimensions
- mask, input_fidelity and partial_images are unsupported, and any moderation value returns HTTP 400
- Accepts only the image_generation tool; function tools and the web_search tool return HTTP 400
Sources
Last verified Page updated 2026-10-13