Meta Muse Video: preview only, so what can you use for Reels?
Meta's July 7, 2026 post previews Muse Video with native audio and says it is coming soon. No API details yet. How to check what Sume lists for 9:16 video.
Muse Video is Meta's video generation model, shown as an early preview on July 7, 2026, and the post says it is coming soon to creators and in Meta AI. The post gives no resolution, frame rate, length or API details, so there is nothing to build a Reels pipeline on yet.
Source: Meta AI's Muse announcement, read 2026-10-02.
What Meta says
Everything below is Meta's wording, not a measurement.
| Item | Meta's post |
|---|---|
| Date | July 7, 2026 |
| Status | Early preview, coming soon to creators and in Meta AI |
| Base | Same pretraining base as Muse Image |
| Audio | Native audio support |
| Watermark | Plans to extend Content Seal to video soon |
| Ranking claim | No. 3 in human-preference Elo for text-to-video on Arena as of July 5, 2026 |
| Resolution, length, API | Not given |
Why the Content Seal line matters
Content Seal is the invisible watermark Meta currently uses for images. If it reaches Muse Video, clips made with it would carry a marker Meta can read. For Reels you publish yourself through the API, Instagram's changelog describes an is_ai_generated field for self-disclosure, covered in our post on that parameter. A watermark and a disclosure flag are separate things.
What to use today on Sume
Sume's docs do not list Muse Video. To see what is available, call GET /v1/videos/models. Each model entry reports supported_aspect_ratios, supported_durations and supported_resolutions, so you can filter for 9:16 before you spend anything.
The docs describe one example of a 9:16 model: Gemini Omni Flash 1.1 at 3 to 10 seconds with native synced audio. Other models have their own limits, so read the catalog entry, not this post.
What to do now
Three practical steps.
- Do not plan a campaign on a preview. Meta has not published limits.
- Keep your Reels pipeline model-agnostic: aspect ratio, length and audio are the inputs that matter.
- Re-read Meta's page when it moves from preview to release.
Sources
Related posts
More in Models
- MiniMax H3 API rate limit: 300 RPM, 30 in-flight vs Sume
MiniMax lists 300 requests per minute and 30 in-flight tasks for H3, and 20 RPM for Hailuo. Sume queues by plan concurrency instead. Read both limits here.
- MiniMax H3 camera prompts: lens, movement, exposure wording
fal's H3 prompting guide says H3 reads film vocabulary: lens, rack focus, handheld, grain. Wording that works as a prompt and a Sume request that sends it.
- MiniMax H3 Max Recast API: swap people in a video, fal price vs Sume
H3 Max Recast swaps people in a source video for reference photos, keeping motion, cuts and audio. fal lists $0.30 a second at 768p; what Sume accepts.
- MiniMax H3 reference inputs: 12 files combined, not 9 + 3 + 3
MiniMax H3 allows 9 images, 3 videos and 3 audio clips as references, but a 12-file combined cap applies. How to stay inside it, and how Sume counts.
Written by Sume