Fei-Fei Li's WorldLabs releases Atlas, a multimodal world model for spatial simulation
量子位 · wechat · 2026-09-02
Fei-Fei Li's WorldLabs has released Atlas, a new multimodal world model hailed as the "first of its kind." Atlas is a multimodal autoregressive diffusion Transformer with core capabilities including:
- Pixel-level camera control generation: Generates up to 1-minute, 1440p controllable video from a single image.
- 3D spatial reconstruction: Reconstructs real-world scenes from sparse inputs, outperforming specialized open-source models.
- Spatiotemporal simulation: Models space and time, supporting Real-to-Sim workflows for robotics.
- Image generation: Creates images and 360° panoramas from text prompts.
The model aims to bridge the gap from real photos/video to 3D space and robotic simulation environments, marking a critical step for embodied intelligence. Early access is currently open to partners.
Related event: World Labs Launches Atlas, an Omni World Model for Spatial Intelligence(64 posts)→
More from Models
- 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
- Polymarket opens betting market on who releases the next Gemini Argon model — Polymarket · 2026-10-18
- Grok matter-of-factly lists which groups user was racist against, exempting Africans and Latin Americans — teortaxesTex · 2026-10-18
- A quick test for benchmaxxed demos: swap in Mirror's Edge and see if it generalizes — teortaxesTex · 2026-10-18
- Did OpenAI distill DeepSeek's gains? Benchmark shows OpenAI's Luna scoring just 38 — teortaxesTex · 2026-10-18