ByteDance’s world model has reportedly debuted, generating navigable 3D worlds from a single sentence.

Dongcha Beating AI News Flash: At the "2026 Wonderful Night" co-hosted by Douyin and Beijing TV, Pan Tianhong (Tim), founder of Film Storm, demonstrated a world model capable of generating explorable virtual environments from text descriptions. During the on-site showcase, users can roam through the generated scenes and input additional commands to modify the visuals. Bloomberg reported in September this year that ByteDance is developing a real-time spatial video model based on Seedance, with fo

@tonmelonfest