Puffin-World: Unified Multimodal World Model with Native 3D State, Open Sourced

Researchers at Nanyang Technological University released and open-sourced Puffin-World, a unified multimodal world model that natively represents 3D world state, jointly modeling physics, geometry and appearance. It supports physically consistent 3D world generation, reconstruction and closed-loop exploration from a single image.

2026-09-03 ~ 2026-09-04 · 3 related posts