Via searchenginejournal.com
Meta's first video generation model with native audio debuts as the company races to compete in the generative AI content creation space
Meta just pulled back the curtain on Muse Video, its first video generation model that can create clips complete with synchronized audio and music. The model, announced on July 7 through Meta’s AI blog as an “early preview,” is currently in closed beta, which means most people can look but not touch. At least not yet.
The release positions Meta squarely in the ring with OpenAI’s Sora, Google’s Veo, and a growing roster of AI video generators that have been steadily improving over the past year. The difference Meta is betting on: native audio that actually syncs with what’s happening on screen.
What Muse Video actually does #
Muse Video was developed by Meta Superintelligence Labs, the division Mark Zuckerberg has been building out as Meta’s answer to the AI arms race. The model generates video with built-in audio capabilities, meaning it doesn’t just create silent clips that need a soundtrack bolted on afterward. Music, sound effects, and audio are generated natively alongside the visuals.
The model also claims competitive performance in three areas that matter most to anyone actually trying to use these tools: prompt adherence (does it make what you asked for), visual quality (does it look good), and temporal consistency (does the subject’s face stay the same from frame to frame instead of melting into something nightmarish). Meta says Muse Video specifically targets improvements in audio-video synchronization and realistic motion rendering.
The broader Muse family #
Muse Video doesn’t exist in isolation. It shares its pretraining foundation with Muse Image, a companion model that launched at the same time. Muse Image is already available in Meta’s consumer apps, including Instagram and WhatsApp, where users can generate images directly within the platforms they already use daily.
Meta also referenced integration with something called Muse Spark, which appears focused on agentic functionalities. In plainer terms, that means AI that doesn’t just respond to a single prompt but can take a series of actions on a user’s behalf.
Meta has promised that Muse Video will be available “soon” to creators and Meta AI users, though no specific date has been announced.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our