NVIDIA Launches Open-Source Project Lyra, Enabling AI-Powered 3D World Generation From a Single Image
Summary
NVIDIA's Spatial Intelligence Lab launches Project Lyra, a groundbreaking open-source series of generative 3D world models capable of creating explorable, 3D-consistent environments from a single image, with Lyra-2.0 now live and the GitHub repository already gaining nearly 1,800 stars.
Key Points
- NVIDIA's Spatial Intelligence Lab releases Project Lyra, an open-source series of generative 3D world models, with Lyra-2.0 launching on April 15, 2026, enabling explorable, long-horizon, 3D-consistent world generation.
- Lyra-1.0, released in September 2025 and accepted at ICLR 2026, introduces feed-forward 3D and 4D scene generation from a single image or video using video diffusion model self-distillation.
- The repository, now boasting 1.8k stars and 164 forks on GitHub, is released under the Apache 2.0 License and provides official implementations for both Lyra versions, with models leveraging Gaussian-based 3D diffusion techniques.