Quick answer: Meta’s Muse family does not have one shared availability status. Muse Image is released on named Meta surfaces. Muse Spark 1.1 is available in Meta AI and through a public-preview Meta Model API. Muse Video is an early preview described as coming soon.
That lifecycle split is the first decision to make. It prevents an image-model launch, an agent-model API preview, and a video research preview from being treated as interchangeable product access.
Muse family at a glance
| Model | Primary role | Verified July 2026 lifecycle | Best next source |
|---|---|---|---|
| Muse Image | Image generation, editing, multi-reference composition, tool use | Released on named Meta surfaces | Meta’s July 7 launch article |
| Muse Spark 1.1 | Multimodal reasoning, agents, computer use, coding | Available in Meta AI; Meta Model API in public preview | Meta’s July 9 launch article and current developer docs |
| Muse Video | Video generation with native audio | Early preview; coming soon | Meta’s July 7 preview and any later release update |
None of these Meta lifecycle statements establishes current Flowith availability.
Choose Muse Image for released image workflows
Meta describes Muse Image as an agentic image model. It can generate and edit images, compose from multiple references, use search and coding tools, and self-refine a draft.
Its launch surfaces were:
- The Meta AI app and meta.ai.
- Instagram Stories in the US.
- WhatsApp in limited countries.
- Facebook, described as coming soon.
These are dated rollout facts, not a promise of identical access in every country or account. The Muse Image model page keeps the generation, editing, Content Seal, and access boundaries together.
Choose Muse Spark 1.1 for agentic reasoning
Muse Spark 1.1 is a different kind of model. Meta positions it for planning, external tools, computer use, coding, multimodal understanding, and multi-agent orchestration.
Meta says it can actively manage a one-million-token context window, retrieve earlier actions, and compact context while preserving critical steps. Those capabilities still need real-workflow evaluation: a large window does not prove that the right detail will survive compaction or reach the right tool.
Developer access launched through the Meta Model API public preview. Public preview is a meaningful boundary. Check current developer sources for endpoint shape, limits, pricing, data policy, compatibility, and production commitments.
Use the Muse Spark 1.1 model page for the source ledger and implementation questions.
Treat Muse Video as preview research
Muse Video was not announced as a released product in the verified source. Meta called it an early preview and said it was coming soon to creators and Meta AI.
The preview includes native audio, but Meta also named work still in progress:
- Audio-video synchronization.
- Physically accurate fast motion.
- Extension of Content Seal to video.
Those gaps matter more than a dated leaderboard rank when deciding whether a production workflow is ready. The Muse Video preview page keeps them visible.
Do not collapse these five boundaries
1. Meta AI access vs Meta Model API access
An experience available in the Meta AI app does not automatically have a callable developer API. For Muse Spark 1.1, Meta explicitly named a separate public-preview API.
2. Released vs coming soon
Muse Image’s named launch surfaces do not prove that Muse Video is available. The two models appeared in one article with different lifecycle language.
3. Model capability vs account entitlement
A launch article can establish model behavior without establishing access for a particular country, plan, account, or workspace.
4. Provenance feature vs universal coverage
Meta says Content Seal applies to Muse Image outputs made in the Meta AI app and on meta.ai. It said extension to video was planned. Verify the exact surface and output path rather than assuming universal coverage.
5. Meta availability vs Flowith availability
No reviewed source establishes a Flowith model mapping for Muse Image, Muse Video, or Muse Spark 1.1. Use the live Flowith workspace as the current availability source.
A practical evaluation plan
- Pick the exact lifecycle: released product, public-preview API, or research preview.
- Recheck the current first-party page for geography, account, and API changes.
- Use a representative task: a multi-reference edit for Image, a tool-and-context workflow for Spark, or an audio-and-fast-motion brief for Video after release.
- Define review criteria before running the task.
- Record where output provenance, permissions, pricing, and terms are actually documented.
- Keep model-family links reciprocal so readers and crawlers can move between task and lifecycle intent.
Frequently asked questions
Is Meta Muse one model?
No. The verified launch sources describe Muse Image, Muse Video, and Muse Spark 1.1 as different models with different tasks and lifecycle states.
Which Muse model is released?
Muse Image launched on named Meta surfaces. Muse Spark 1.1 launched in Meta AI and through a public-preview Model API. Muse Video was described as an early preview coming soon.
Does Muse Image have an API?
The reviewed July 7 launch article does not establish a Muse Image API. Check current Meta developer sources rather than inferring API access from Meta AI product availability.
Can developers access Muse Spark 1.1?
Meta launched developer access through the Meta Model API in public preview. Verify current preview requirements, limits, pricing, and terms before implementation.
Is Muse Video available now?
The verified source says coming soon, not released. Check for a newer official Meta announcement before treating it as available.
Are Meta Muse models available on Flowith?
This guide does not claim that. Meta product and API availability do not establish a Flowith model mapping.
Primary sources
- Introducing Muse Image and Muse Video — Meta’s July 7, 2026 image launch and video preview.
- Introducing Muse Spark 1.1 — Meta’s July 9, 2026 model and public-preview API announcement.