Movie Gen is a set of Meta research models, described in 2024, that generate HD video with matching sound from text and can edit or personalise videos.
In plain words
Movie Gen is several models working together. The main video model is a 30-billion-parameter transformer that makes 1080p clips of up to 16 seconds in various shapes. A separate audio model adds synchronised sound effects and music. The system can also edit an existing video from a text instruction, or put a specific person into a new video from a single photo of them. Meta published a public benchmark so others can compare their results to it.
A quick example
Imagine a text prompt asking for a paper boat drifting down a rainy street. Movie Gen would produce the clip, then its audio model could add the patter of rain and a matching soundtrack.
Where you'll meet it
- Video generation research and benchmarking.
Related
Sources
- PaperMovie Gen: A Cast of Media Foundation Models, Polyak et al. (arXiv) · read 29 Sept 2026
- DocsMovieGenBench, Meta (GitHub) · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.