Status: previewed · not yet released · updated 2026-07-12
Meta previewed Muse Video — its AI video generator with native audio — on July 7, 2026, and it is not publicly available yet. A wave of sites already claims to run it; none of them do. This is the independent tracker: what the Muse video generator actually is, when you can use it, and what to create with today.
The three models in Meta's Muse family, and where each one actually stands.
Previewed — not yet released
Announced July 7, 2026 as a preview. Built on the same pretraining base as Muse Image, with native audio support. Meta says it is "coming soon to creators and Meta AI" — no date, regions, pricing or API announced.
Live (US, inside Meta products)
Launched July 7, 2026. Agentic image generation and editing — it searches the web, uses tools and refines its own output. Debuted at No. 2 on the Arena human-preference leaderboard for text-to-image and image editing.
Live + developer API (US preview)
The family's language model, released July 9, 2026 alongside the new Meta Model API (public preview for US developers). The only Muse model with an API so far — text only.
Muse Video is the upcoming AI video generator from Meta Superintelligence Labs — the lab's first video model, previewed alongside the now-live Muse Image. Meta has confirmed three things: it's built on the same pretraining base as Muse Image (so expect a consistent visual style across the family), it generates native audio with the video, and it delivers what Meta calls "exceptional visual fidelity."
That native-audio detail matters. It puts the Muse video generator in the same class as OpenAI's Sora 2 and Google's Veo 3 — the current generation of AI video generators that produce sound and picture together — rather than the silent-clip tools of earlier years. What nobody outside Meta knows yet: clip length, resolution, pricing, regions, or whether a Muse Video API will exist. Meta's only timing language is "coming soon to creators and Meta AI."
After days of backlash, Meta withdrew the ability to prompt Muse Image directly on Instagram accounts. The broader Instagram-grounding feature — drawing on public content for social context — still ships.
Meta released the family’s language model and opened the Meta Model API in public preview for US developers. It serves Muse Spark only — neither Muse Image nor Muse Video is available through it.
Reports confirmed Muse Image can reference public Instagram photos without notifying the people in them. Private accounts are excluded by default; public accounts can opt out via Settings → Sharing and Reuse.
Meta Superintelligence Labs shipped its first media models. Muse Image went live in the Meta AI app, meta.ai and Instagram Stories (US). Muse Video was previewed — same pretraining base, native audio, "coming soon to creators."
| Model | Maker | Status | Audio | Access | API |
|---|---|---|---|---|---|
| Muse Video | Meta (MSL) | Preview — unreleased | Native audio (confirmed) | Nobody outside Meta yet | Not announced |
| Sora 2 | OpenAI | Available | Native audio | Sora app / sora.com | Yes (OpenAI API) |
| Veo 3 | Google DeepMind | Available | Native audio | Gemini / Flow | Yes (Gemini API / Vertex) |
| Kling / Seedance class | Kuaishou / ByteDance | Available | Varies by model | Web apps + aggregators | Yes (via aggregators) |
Until Meta ships it, a Muse video generator simply does not exist outside Meta — every clip you can make today comes from Sora 2, Veo 3 or the value-tier models. What Muse Video changes is distribution: Meta can put an AI video generator inside Instagram and WhatsApp overnight.
While the Muse video generator stays unreleased, the image side of this wave is fully usable. Muse Image itself is US-only and locked inside Meta products — but the models it competes with on the Arena leaderboard are open to everyone: run Nano Banana 2, Seedream 4.5 and GPT Image 2 side-by-side on Yix — free starting credits, no Meta account. For video right now, Sora 2 and Veo 3 are the available benchmarks (full guide).
No. Muse Video was previewed by Meta on July 7, 2026 and has not been released — no app, no waitlist, no API. Any website claiming to be a working Muse video generator today is serving a different model under the Muse name.
Meta has only said "coming soon to creators and Meta AI" — no date, regions or pricing. We track every confirmed update on our release-date page.
Muse Video is the upcoming AI video generator from Meta Superintelligence Labs, built on the same pretraining base as Muse Image, with native audio support and 'exceptional visual fidelity' as the headline pitch.
Not yet. Muse Image is live in the Meta AI app, meta.ai and Instagram Stories in the US, and WhatsApp in limited countries. A Meta account is required everywhere, and Meta has not given an international timeline.
Only Muse Spark (the text model) has an API — the Meta Model API, in public preview for US developers. Muse Image and Muse Video have no public API, weights or third-party integrations.
For video with native audio, Sora 2 and Veo 3 are the available benchmarks. For the image side of the Muse family, top-ranked alternatives like Nano Banana 2, Seedream 4.5 and GPT Image 2 are usable from any country without a Meta account.
No. MuseVideoGenerator.com is an independent tracker and guide. Muse, Muse Image, Muse Video and Meta AI are trademarks of Meta Platforms, Inc.
“Every 'Muse video generator' site I found claimed to already run it — this is the only one that told me straight it isn't out yet, and what to use instead.”
“The API status table saved me an afternoon — I was about to build on a "Muse API" reseller that turned out to be proxying something else entirely.”
“I check the release tracker every few days. Clear timeline, no hype, and the Instagram opt-out steps actually worked for our brand account.”
“The Sora vs Veo vs Muse comparison is the clearest one I have seen — bookmarked it for the day Muse Video actually ships.”