World Labs
@drfeifei
Today we are introducing Atlas, our next-generation world model. Atlas is an omni model that we pretrained from scratch to natively operate on text, images, video, and 3D. It is a multimodal autoregressive diffusion transformer: all inputs are combined into a shared spatial context.
Atlas can perform a broad range of tasks spanning world generation, reconstruction, and simulation: camera-controlled generation of up to 1 minute of video at 1440p; spatial reconstruction from one to dozens of images with explicit 3D outputs; space-time simulation for VFX and Real-to-Sim robotics; and image and 360 panorama generation.
Atlas is built to scale, and will power future versions of Marble and other World Labs products. Early access is opening in the coming weeks.
