Meta previews AVM world model: volumetric video from synced cameras

Scobleizer · x · 2026-11-20

Meta has given a first look at AVM, its new "world model" capable of creating volumetric video from multiple synchronized cameras without relying on 4DGS. No source code, inference performance, or inference hardware details were provided.

Related event: Meta Unveils AVM, a World Model Generating Walkable 3D Scenes Without 3D Data(4 posts)→

Original post →

More from Multimodal

Multimodal channel →