Meta/muse-image-1-0/Released Aug 2026

Muse Image 1.0

Fast image generation and precise editing

Commercial use

Image

About Muse Image 1.0

Muse Image 1.0 is Meta's image model for generation, editing, and multi-image composition. It takes an agentic approach to a request: it plans the image first, pulls real references from the web, and generates and runs code where precision matters before rendering, with search grounding and reasoning included in the per-image price.

When editing, the model is designed to change exactly what is specified and leave everything else untouched. It edits or composes one or more reference images, generates from text alone, and keeps a series consistent when generations share the same references.

See what this model can do

A handful of real prompts and their outputs.

A field-guide plate of five North American woodpeckers drawn to scale with common and scientific names.

Natural History

Woodpecker Comparison Plate

"Researches each species to get plumage, relative size, and current scientific names right."

Looking up the nave of the Sagrada Familia, with branching columns and orange and green light from stained glass.

Architecture

Sagrada Familia Nave

"Uses real references to reproduce the branching columns and star-patterned vault of a real building."

Macro of a salmon-pink enamel watch dial with sunburst and hobnail guilloche engraving and blued hands.

Macro Product

Guilloche Watch Dial

"Keeps two engraved guilloche patterns regular and sharp at extreme magnification."

A blueprint-style cutaway of the Apollo Lunar Module with ten labeled systems and a title block.

Technical Diagram

Apollo LM Cutaway

"Grounds the spacecraft geometry in NASA references and places every label on the right part."

A museum placard with a line chart of atmospheric CO2 at Mauna Loa rising from about 320 to about 420 ppm.

Data Graphic

Keeling Curve Placard

"Plots real measurement data with the seasonal cycle, clean axes, and a readable caption."

Overhead macro of a sunflower head with seeds in interlocking golden-angle spirals and a honeybee at the edge.

Nature Macro

Golden-Angle Sunflower

"Computes the seed pattern so the florets form true interlocking Fibonacci spirals."

Science Illustration

The Ten Cloud Genera

"Places each cloud type at its correct altitude with one accurate label each."

A glowing Gothic rose window with twelve petals and twelve outer roundels in ruby, cobalt, and gold glass.

Symmetry

Twelve-Fold Rose Window

"Uses generated geometry to hold exact twelve-fold symmetry across the stone tracery."

Natural History

The Eight Bears

"Researches all eight species to get relative size, markings, and scientific names right."

Ready to build with Muse Image 1.0?

Try Muse Image 1.0 in the Workbench to prompt it, compare outputs, and iterate on prompts without writing any code. When you're ready to ship, call the same model from our API and build your own apps on top of it.

curl -sSf -X POST https://hub.oxen.ai/api/ai/images/edit \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer $OXEN_API_KEY" \
    -d '{
  "model": "muse-image-1-0",
  "prompt": "A watercolor painting of a red fox curled up under a birch tree in first snow, soft morning light, muted palette.",
  "aspect_ratio": "16:9",
  "thinking_level": "high",
  "output_format": "png"
}'

Request parameters

Every field Muse Image 1.0 accepts in the request body. See the API reference for response formats.

POSThttps://hub.oxen.ai/api/ai/images/edit

Field	Type	Required	Default	Description
`prompt`	`string`	Yes	—	Text description of the image to generate, or of the edit to apply to the reference images.
`input_image`	`array<string>` nullable	No	—	Optional reference image(s). Pass one to edit it, several to compose them, or leave empty to generate from the prompt alone.
`aspect_ratio`	`string`	No	`"16:9"`	`auto1:116:99:164:33:43:22:321:99:21`
`thinking_level`	`string`	No	`"high"`	How much the model reasons before generating. 'high' improves complex compositions at higher latency; 'low' renders faster. `lowhigh`
`output_format`	`string`	No	`"png"`	File format for the generated image. `pngjpegwebp`

Model JSON Image editing reference Async queue Models endpoint