Choose Auto, TokenLab Verified, or Official for each request, with prices shown up front. See what's new

Meta: Muse Image

Muse Image is Meta's flagship image generation model from Meta Superintelligence Labs. It is built for prompt-faithful image creation, precision editing, and multi-reference composition, with strong text rendering and the ability to refine existing photos through localized markup-based edits. Meta positions it as an agentic image model that plans layouts, uses search and coding tools to improve accuracy, blends multiple visual references intelligently, and handles both creative generation and practical visual tasks such as infographics, QR codes, restorations, product-style mockups, and photobomber removal.
Compare models
muse-image
AvailableMetaImageSync
Price
From $0.01
Modalities
Image

About Muse Image

Muse Image is Meta's image generation model, the first from Meta Superintelligence Labs, built for prompts that need planning rather than a quick render. It pairs with the Muse Spark language model to lay out a composition, look up context when a request needs it, and blend several reference photos. It also edits existing pictures, including changes marked directly on the image, and renders legible text for infographics and signage.

Where it works well

  • It reasons about a request before drawing, planning layout for complex prompts such as infographics, guides and multi-element scenes.
  • Text inside images comes out clean, which suits posters, diagrams and mockups with readable labels.
  • Editing accepts annotations: circling or sketching on a photo tells the model exactly which region to change.
  • Several reference photos can be combined into one coherent picture, such as a person placed in front of a landmark.
  • Restoration and clean-up tasks, like repairing old photos or removing a bystander, are named uses in Meta's announcement.

When to choose another model

  • It is positioned as an agentic system, so requests that involve planning or lookups can take longer than a single-pass image model.
  • Meta withdrew its feature for referencing public Instagram accounts in prompts, so people-based references rely on photos you supply.
  • It covers text-to-image and image editing only; it does not generate video or audio.

Getting started

  1. Create API key

    Create a key in Console, then use it with every model on the platform.

  2. Send your first request

    Copy the example for your language and run it against the endpoint.

    Image editingText to imageTokenLab endpoint
    POST/v1/images/generations
    curl -X POST "https://api.tokenlab.sh/v1/images/generations" \
      -H "Authorization: Bearer sk-xxx" \
      -H "Content-Type: application/json" \
      -d '{
      "operation": "image-to-image",
      "model": "muse-image",
      "prompt": "A minimalist product photo of matte black headphones on a soft blue background.",
      "image_url": "https://example.com/source.png"
    }'

Parameters

Image editing

ParameterValues and limitsDefaultNotes
height1008, 1152, 1280, 1344, 1600, 1792, 1920, 20481600Use an integer. Use a listed width and height pair: 1600x1600 (2k, 1:1), 1920x1280 (2k, 3:2), 1280x1920 (2k, 2:3), 1792x1344 (2k, 4:3), 1344x1792 (2k, 3:4), 2048x1152 (2k, 16:9), 1152x2048 (2k, 9:16), 2352x1008 (2k, 21:9).
imageNot providedNot providedAlias of the combined images input. See input_limits.images for the total input image count.
image_urlNot providedNot providedAlias of the combined images input. See input_limits.images for the total input image count.
image_urlsNot providedNot providedAlias of the combined images input. See input_limits.images for the total input image count.
imagesInputs: ≥ 1, ≤ 10Not providedLimits apply when this input is supplied. Image aliases are combined into one input image list; the limit applies to the combined list.
n11Use an integer.
operationRequiredNot providedNot providedNot provided
output_formatjpeg, png, webpNot providedDefault encoding is shared by these equivalent values: jpg, jpeg.
promptNot providedNot providedMinimum length: 1 characters. Maximum length: 200000 characters.
width1152, 1280, 1344, 1600, 1792, 1920, 2048, 23521600Use an integer. Use a listed width and height pair: 1600x1600 (2k, 1:1), 1920x1280 (2k, 3:2), 1280x1920 (2k, 2:3), 1792x1344 (2k, 4:3), 1344x1792 (2k, 3:4), 2048x1152 (2k, 16:9), 1152x2048 (2k, 9:16), 2352x1008 (2k, 21:9).

Text to image

ParameterValues and limitsDefaultNotes
height1008, 1152, 1280, 1344, 1600, 1792, 1920, 20481600Use an integer. Use a listed width and height pair: 1600x1600 (2k, 1:1), 1920x1280 (2k, 3:2), 1280x1920 (2k, 2:3), 1792x1344 (2k, 4:3), 1344x1792 (2k, 3:4), 2048x1152 (2k, 16:9), 1152x2048 (2k, 9:16), 2352x1008 (2k, 21:9).
imagesNot providedNot providedText-to-image does not accept source images. Do not send this field for that operation. Use one of this model's declared image operations for source-image requests.
n11Use an integer.
operationRequiredNot providedNot providedNot provided
output_formatjpeg, png, webpNot providedDefault encoding is shared by these equivalent values: jpg, jpeg.
promptNot providedNot providedMinimum length: 1 characters. Maximum length: 200000 characters.
width1152, 1280, 1344, 1600, 1792, 1920, 2048, 23521600Use an integer. Use a listed width and height pair: 1600x1600 (2k, 1:1), 1920x1280 (2k, 3:2), 1280x1920 (2k, 2:3), 1792x1344 (2k, 4:3), 1344x1792 (2k, 3:4), 2048x1152 (2k, 16:9), 1152x2048 (2k, 9:16), 2352x1008 (2k, 21:9).

Fields outside this contract are rejected.

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

Image editing

per image
Official price
$0.01
Verified price
—
Discount
—

Text to image

per image
Official price
$0.01
Verified price
—
Discount
—

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours
Requests
Success rate
P95 latency
Total tokens
Model performance
Metrics appear once privacy and data volume thresholds are met.

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open Muse Image in Console with a prompt ready to edit or send.

Help me create an image with muse-image using image-edit at /v1/images/edits. Show the result, cost, and any limits I should know.

Use cases

Best for
  • Text to Image
  • Infographics and explainers

    Describe the data and the structure, and let the model plan the layout and set the labels in readable type.

  • Photo restoration

    Upload a damaged or faded photo and ask for repair, with the edit applied to the existing picture instead of a new image.

  • Composite scenes

    Supply two or three reference photos, such as a selfie and a travel shot, and ask for one believable combined image.

  • Marked-up retouching

    Circle an object in a photo and instruct its removal or replacement while the rest of the frame stays put.

Prompt examples

Create an infographic explaining how a heat pump works, four labeled steps, flat illustration style, readable captions.

Restore this scanned 1970s family photo: remove the creases, fix the faded colors, keep the faces unchanged.

Combine these two photos so the person stands in front of the lighthouse at sunset, matching the lighting.

FAQ

What is Muse Image?

It is Meta's image generation model from Meta Superintelligence Labs. It builds on the Muse Spark language model to understand complex prompts and visual references, and it can plan a layout before it renders the picture.

Can Muse Image edit existing photos?

Yes. It supports edits to uploaded photos, including removing people or objects and changing style. Meta describes drawing or annotating directly on the photo to point at what should change.

Does it render readable text?

Meta highlights clean text rendering inside images, for example in infographics and guides. Keep on-image text short and spell it out in the prompt for the most reliable results.

Can I give it several reference images?

Yes. It is designed to blend multiple visual references into one composition, so you can combine subjects, places and styles from separate photos in a single request.

Why is it called agentic?

Meta says the model can use tools such as web search and code, and refine its own output, before returning the image. This helps with factual or structured pictures such as QR codes and infographics.

How much does Muse Image cost?

On TokenLab, Muse Image costs $0.01 per image. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

Which endpoint should Muse Image use?

Use https://api.tokenlab.sh/v1/images/edits for Muse Image. The request example below shows the matching code shape.

Which operations does Muse Image support?

Muse Image supports Image editing, Text to image. Select an operation above to see its endpoint and request example.

Compare Muse Image

Guides that use Muse Image

Sources

Reviewed Oct 2, 2026

Related models