World Labs Unveils Atlas: Omni World Model for 3D Reconstruction and Simulation
venturetwins · x · 2026-09-02
World Labs has released Atlas, a next-generation omni world model for spatial intelligence. It is a native multimodal autoregressive diffusion Transformer that processes text, images, video, and 3D data within a shared spatial context.
Key Capabilities:
- Camera-Controlled Generation: Generates images and videos from inputs with pixel-perfect camera control, outputting up to 1 minute of 1440p video.
- Spatial Reconstruction: Reconstructs real-world scenes from 1 to dozens of input images. It generates novel view frames and explicit 3D outputs, outperforming state-of-the-art specialized models.
- Space-Time Simulation: Models space and time from input videos for reframing effects and Real-to-Sim workflows for robotics.
- Image Generation: Creates images and 360 panoramas from text, handling complex prompts, text rendering, and various visual styles.
Atlas demonstrates strong scaling properties, with performance improving as training compute increases.
Related event: World Labs Launches Atlas, an Omni World Model for Spatial Intelligence(64 posts)→
More from Models
- Mocked benchmark shows DeepSeek's document chart grounding badly misaligned — burkov · 2026-10-18
- Dev complains Claude's habit of saying "not done" is his biggest pain point — jasonkneen · 2026-10-18
- Open-weight TYGER speech decision model makes 80ms turn-taking calls for full-duplex voice agents — solyarisoftware · 2026-10-18
- Paying Google AI subscriber blocked: custom-domain Workspace accounts can't access AI Studio — ChrisRackauckas · 2026-10-18
- Running Ling 3.1 Flash (560B MoE, 25B Active) Across 4 DGX Sparks via llama.cpp RPC — BerryBlush_Rogers691 · 2026-10-18
- Users press OpenAI on long-delayed Adult Mode a year after it was promised — ___Patrice___ · 2026-10-18