World Labs releases Atlas, an omni world model native to text, images, video, and 3D

theworldlabs · x · 2026-09-24

World Labs has introduced Atlas, its next-generation omni world model for spatial intelligence, pretrained from scratch to natively operate on text, images, video, and 3D. Architecturally it is a multimodal autoregressive diffusion transformer: all inputs combine into a shared spatial context, and generation stays 3D-consistent with everything seen.

Key capabilities:

Camera-controlled generation: pixel-perfect camera control from one or more images, up to 1 minute of video at 1440p;
Spatial reconstruction: reconstructs real scenes from 1 to dozens of images, producing novel-view frames and explicit 3D outputs, outperforming specialized 3D-reconstruction SOTA;
Space-time simulation: models space and time from video, enabling reframing and Real-to-Sim workflows for robotics;
Image generation: text-to-image and 360 panoramas with complex prompt following and text rendering.

Atlas scales with training compute, a trend the team expects to hold. It will power future versions of Marble; beta signups are open.

Original post →

World Labs releases Atlas, an omni world model native to text, images, video, and 3D

More from Multimodal