Atlas
Model ● Covered in 7 stories + Follow
Atlas (by World Labs) is described as a multimodal “omni model” for spatial intelligence that reconstructs and renders 3D scenes from a small set of photos or a single image input, with camera control and the ability to generate 1440p video up to one minute. Recent coverage says it has been made available in early access for select enterprises and supports text, images, video, and 3D for robotics-oriented scene simulation. Atlas is also mentioned as an AI agent used alongside other agentic components in production systems run through platforms like monday.com on Amazon Bedrock.
Updated 11 September 2026
Specifications
No specifications recorded yet.
Latest developments
Q3 2026
AWS Bedrock enables multi-model AI architectures for enterprise applications Deployment
- The Sequence Learning Loop - Issue 929: Learn About Meta Muse Spark, World Labs’ Atlas and Gemini 3.8 Flash
- Fei-Fei Li’s World Labs debuts Atlas, a world model showcase for advanced spatial intelligence
- Atlas by World Labs
- AI Teammates: how monday.com runs production AI agents on Amazon Bedrock
Q1 2026
Q4 2025
Relationships
Products & technology
- Integrated with ChatGPT · 1 source
- Integrated with OWL · 1 source
- OWL integrated with this model · 1 source
- Together AI develops this model · 1 source
- World Labs Inc. develops this model · 1 source
- World Labs develops this model · 1 source