News confirmed medium confidence

Meta Previews Muse Video With Native Audio and a Vendor-Cited #3 Arena Rank — No Public Access Yet

Announced July 7 alongside Muse Image, Meta's video model generates synchronized audio but ships with no public product, pricing, or API — and its Arena ranking is Meta's own citation.

Meta previewed Muse Video on July 7, 2026 — the same day it launched Muse Image — describing a text-to-video model with natively generated audio, according to explainers from aireiter and musevideo.me citing Meta's announcement.

Context

Muse Video shares its pretraining base with Muse Image and extends it to video with synchronized, natively generated sound. Meta cites a #3 rank on the Arena text-to-video human-preference Elo leaderboard as of about July 5 — a vendor-cited snapshot, not an independent verification. Some early trackers dated the announcement July 8; Meta's official post is dated July 7, with broader coverage following July 8.

What changed — and what did not

Why it matters

The preview-not-product framing matters because Muse Image shipped publicly the same day and the two are easily conflated. Strategically, Meta is signaling it intends to compete in generative video — with native audio as the differentiator — just as OpenAI exits the category (Sora's API carries a September 24 shutdown date) and Google opens Gemini Omni Flash. But until access opens, every capability claim is Meta's own.

Limitations

Confidence is high on the preview's existence and status; medium on the Arena ranking, which is vendor-cited and snapshot-dependent. No independent evaluation is possible without public access.

Sources

*Last reviewed July 22, 2026. Next checkpoint: creator and Meta AI availability, date unannounced.*

Update note: Last reviewed 2026-07-22. Next checkpoint: monitor official channels and the linked source record.

Sources

Drafted with AI assistance from source briefs; reviewed for citation completeness and label accuracy.