Generate and edit
Meta describes instruction-following generation, precise edits, iterative refinement, and composition from multiple image references.
Meta Superintelligence Labs / Muse Image / released July 7, 2026
Quick answer
Muse Image is released. Its tools are part of the model story.
Meta launched Muse Image for generation, editing, multi-reference composition, search and coding tool use, and self-refinement across named Meta surfaces.
What the official source supports
Use tools
The launch article says Muse Image can use search for current visual context and code for artifacts such as plots, QR codes, GIFs, and interactive media.
Self-refine
Muse Image can revise a local detail, regenerate a larger section, or switch tactics when a draft does not meet the task.
Access boundary
Product rollout is surface-specific
The July 7 record names the Meta AI app, meta.ai, Instagram Stories in the US, WhatsApp in limited countries, and a future Facebook rollout. It does not establish an API, a uniform global entitlement, commercial rights, pricing, or Flowith availability.
Provenance note
Content Seal has a defined scope
Meta says images made in the Meta AI app and on meta.ai carry its invisible Content Seal signal. Do not generalize that statement to every export path or third-party surface without a current source check.
Muse route map
Muse Spark 1.1 Reasoning and agentic partner with a public-preview Model API. Muse Video Preview-only video sibling with native audio and explicit current gaps. Muse family guide Choose the released, preview, or developer surface without mixing lifecycle states. GPT Image 2 Compare another source-backed image model route.
Muse Image questions, answered
Muse Image is Meta's image generation and editing model launched on July 7, 2026. Meta describes it as an agentic media model that can use search and coding tools, self-refine, and work with multiple image references.
Meta said Muse Image was available in the Meta AI app and on meta.ai, Instagram Stories in the US, and WhatsApp in limited countries, with Facebook access coming soon. Recheck current product availability before relying on this launch snapshot.
Yes. Meta's launch record describes instruction-based editing, coherence across editing turns, and iterative refinement toward a target result.
Meta says Muse Image can compose people, objects, clothing, styles, and environments from multiple input references and can interleave text and images in a prompt.
Content Seal is Meta's invisible provenance watermark. Meta says images created in the Meta AI app and on meta.ai carry the signal, and it previewed a detection tool. Verify coverage on the exact product surface you use.
This page does not claim Flowith availability. Meta's launch establishes access on named Meta surfaces, not a Flowith model mapping; check the live Flowith workspace separately.