World Labs releases Atlas, an omni world model trained from scratch on text, images, video and 3D in one shared spatial context

World Labs·high signalXBlueskyLinkedInCopy link

World Labs introduced Atlas on 2026-09-01, a multimodal autoregressive diffusion transformer that takes text, images, video and 3D as native input types and generates what comes next while staying 3D-consistent with everything already seen. Concrete capabilities include camera-controlled generation of up to one minute of 1440p video from one to six reference images with explicit camera geometry as an input type rather than a text instruction, spatial reconstruction from one to dozens of images that outputs explicit 3D and beats specialized reconstruction models, and space-time simulation for real-to-sim robotics workflows. Atlas will back future versions of Marble and is currently early access only.