01
In brief
Meta's Muse Image is the first image-generation model from Meta Superintelligence Labs. The model originally launched inside Meta AI in July 2026, and on August 26 it became newly accessible to developers through Vercel AI Gateway under the model ID meta/muse-image-1.0.
That API availability changes the practical story around Muse Image: developers can now call the model from applications instead of using it only through Meta's consumer surfaces.
02
What is Muse Image?
Muse Image is Meta's image-generation and editing model. Meta says it was built to follow complex instructions, combine multiple visual references and support iterative editing rather than only one-shot text-to-image generation.
In Meta AI, users can generate new images, upload photos for edits, use suggested presets and continue refining an image conversationally. Meta also uses Muse Image in creative experiences across Instagram and WhatsApp, with additional Meta surfaces and advertiser workflows announced for expansion.
03
Muse Image API availability
As of August 28, 2026, Muse Image 1.0 is available through Vercel AI Gateway. Vercel lists the model as meta/muse-image-1.0 and exposes it through the AI SDK image-generation interface.
The current Gateway documentation shows one Meta provider and supports both generating an image from a text prompt and editing an input image with instructions. Vercel also documents reference-image inputs for steering generated results.
This article does not assume a separate direct Meta developer endpoint unless Meta publicly documents one. The verified developer-access path here is Vercel AI Gateway.
04
Muse Image pricing
Vercel currently lists Muse Image provider output pricing at $0.01 per image. That is the price displayed on the AI Gateway model page at the time of review and can change.
Vercel says AI Gateway reflects provider pricing without a platform markup on inference. Developers should still check the live model page before budgeting production workloads because provider pricing, routing and availability can change quickly.
Meta separately says using Muse Image in Meta AI is free for everyday creation, with additional usage available through Meta subscription plans. Meta's launch post does not publish a simple numeric consumer quota, so this article does not invent one.
05
What can Muse Image do?
Text-to-image generation
You can send a written prompt and generate a new image. Meta highlights instruction following, visual composition and text rendering as core capabilities.
Image editing
Muse Image can take an existing image plus an instruction and return an edited result. The same model handles generation and editing, so developers do not need to switch to a separate editing model in Vercel's current integration.
Multiple references
Meta's technical material emphasizes multi-image composition and reference blending. Vercel's Gateway documentation also supports passing reference images alongside the text prompt to steer the result.
Iterative creative workflows
Inside Meta AI, Muse Image keeps conversational context so users can make follow-up changes without restarting. API applications can build their own iterative experience around repeated image calls, though developers should not assume Meta AI's complete consumer conversation state is automatically reproduced by a single Gateway request.
06
Muse Image in Instagram and Meta AI
Muse Image launched as part of Meta AI and also powers creative experiences in Meta's apps. Meta announced more than 30 AI-powered effects for Instagram Stories and image generation in WhatsApp chats in selected markets, with further rollout planned.
Meta also said Muse Image is coming to additional surfaces including Facebook and Messenger and to advertisers through Advantage+ creative. Availability can differ by country, account and product surface, so 'Muse Image is available' should not be interpreted as every Meta feature being globally enabled.
07
How Muse Image compares conceptually with other image models
Muse Image enters a crowded field that includes OpenAI's GPT Image family, Google's Gemini image models, Black Forest Labs' FLUX family and other specialist image APIs.
The distinctive part of Meta's positioning is the combination of image generation, editing, multiple references and integration with Meta's social context and apps. That does not by itself prove it is better than a competitor for every use case. Cross-model comparisons should use current pricing, output quality, editing behavior, latency and licensing/usage terms rather than a single vendor benchmark.
08
Benchmarks and quality claims
Meta reported that Muse Image held the No. 2 position on Arena for text-to-image, single-image editing and multi-image editing using human-preference Elo rankings around the July launch. Those rankings are time-sensitive and can change as new models enter the leaderboard.
Meta also introduced Content Seal, an invisible provenance signal for images created through Muse Image in Meta AI. Meta says the signal is designed to survive common transformations such as cropping, compression and screenshots. Developers should verify whether identical provenance behavior applies to every third-party API path before making compliance claims.
09
How to call Muse Image through Vercel
Vercel's AI SDK uses the experimental image-generation interface. The basic pattern is to set the model to meta/muse-image-1.0 and call generateImage with a prompt. A separate Project Monet implementation guide covers text generation, editing with an input image and reference-image workflows in more detail.
10
Important limitations and unknowns
- Vercel currently shows one Meta provider for the model, so provider diversity is limited compared with some text models.
- Public API rate limits and every supported image-size/output option are not inferred here unless documented on the current endpoint.
- Meta consumer availability varies by country and product surface.
- Meta's free everyday-creation claim does not specify a universal numeric quota.
- Provider pricing and model availability can change after publication.
11
Who should pay attention?
Muse Image is especially relevant to developers building image-generation tools, product-photo workflows, social creative applications, marketing-asset generators and editing experiences where one model handling both generation and edits simplifies the product architecture.
For creators and marketers, the wider Meta ecosystem matters too: Muse Image is not only an API model but part of Meta's strategy for creation across Meta AI, Instagram, WhatsApp and advertising tools.
12
Frequently asked questions
Is Muse Image available through an API?
Yes. As of August 28, 2026, Vercel AI Gateway exposes meta/muse-image-1.0 for developers.
How much does Muse Image cost?
Vercel currently lists the provider price at $0.01 per output image. Check the live Gateway page before production use because pricing can change.
Can Muse Image edit existing images?
Yes. Vercel documents image editing with an input image and instruction, while Meta also demonstrates conversational photo editing in Meta AI.
Can Muse Image use reference images?
Yes. Meta highlights multi-reference composition and Vercel documents reference-image input through its image-generation API.
Is Muse Image free?
Meta says everyday creation in Meta AI is free, with higher usage available through subscriptions. That consumer statement is separate from developer API pricing.
Is Muse Image available on Instagram?
Muse Image powers some Instagram creative experiences, but availability varies by feature and market. Do not assume the full Meta AI Muse Image interface is present in every Instagram account.
Sources
Primary and supporting sources
Facts were rechecked against the linked sources immediately before publication. Pricing, product availability and rollout status can change.