Puffin-World scales unified multimodal model with native 3D world states
Kang Liao · hf · 2026-09-04
Puffin-World is a unified multimodal framework that jointly models physics, geometry, and appearance via native 3D world states, enabling physically consistent 3D world generation, reconstruction, and closed-loop exploration. Released on Hugging Face.
Related event: Puffin-World: Open-Source Native 3D World Model from NTU(2 posts)→
More from Multimodal
- New style spotted on Midjourney V8.2 with sample renders — azed_ai · 2026-09-04
- Recreating 1997 movie tie-in fast food ads with MiniMax H3 AI video — Certain_Potato_4509 · 2026-09-04
- 1080p AI short generated in 20 steps with MiniMax H3 Hybrid b30 — Phrygian1221 · 2026-09-04
- AI-generated Rolls-Royce with a Daikin AC unit goes viral in Dutch corners — Viralaistudio28 · 2026-09-04
- AI filmmaker beats violence limits in Odyssey short with off-screen sound design — Kyrannio · 2026-09-04
- MiniMax H3 prompt trick: turn one character image into a cinematic digital assembly sequence — Hailuo_AI · 2026-09-04