Hyper3D Launches WorldGen: One Image to an Editable, Physics-Aware 3D Scene in Minutes
量子位 · wechat · 2026-09-01
Hyper3D has released WorldGen, a world-generation model that moves 3D generation from single assets to full scenes: upload one image and get an interactive, editable 3D scene in 2-3 minutes, where every object is an independent asset that can be picked up, moved or swapped, with physics attributes like collision bodies, mass and friction.
The tech builds on CAST, which won Best Paper at SIGGRAPH 2025 (the only commercial winners besides Google and Meta). CAST uses a four-step pipeline—recognize, occlusion-aware completion, place back, physical correction—and the team spent nearly a year fixing cascading errors for production readiness. Applications span embodied AI (closed-loop simulation with D-Robotics and MotrixSim, plus Isaac Sim integration), game level design (assets export to Blender/Unity/Unreal), film (combined with Seedance for controllable camera work), and XR.
Related event: Deemos Releases WorldGen: One Image Generates Fully Editable 3D Worlds(8 posts)→
More from Multimodal
- Flux 3 Text-to-Image Can't Make Realistic Faces at All, Reddit User Reports — EstablishmentFun6629 · 2026-10-18
- MiniMax H3 Ref-to-Video: 15s Shot from 3 Images + Vocals, 1088p Latent Refinement — michael_goldenhour · 2026-10-18
- Dev builds fully local AI tool generating HD 3D textures with Qwen, may sell for $5 — TinfoilTricorn · 2026-10-18
- 14-Agent AI Film Pipeline on One RTX 5090: Fixing Cut Defects Between Clips — developervkmp · 2026-10-18
- 4MP Text-to-Image Comparison: Qwen Image 2.1 Turbo vs MOSS vs Krea 2 vs Flux 3 Klein — FORNAX_460 · 2026-10-18
- Claude writes entire song from just lyrics: "Don't let me scroll" — TinfoilTricorn · 2026-10-18