World Labs, co-founded by Fei-Fei Li, has announced Atlas, a world model that generates, reconstructs, and simulates 3D scenes from as few as one to several dozen images. The company says the omni-model was trained from scratch on text, images, video, and 3D data, and outputs up to one minute of 1440p video. World Labs claims Atlas outperformed specialized models in human-evaluated camera-controlled generation and few-view 3D reconstruction tests, and is available via early access.
No score is assigned. Sources and their independence are shown in the citation chain below.