Viggle releases JST-2, a world foundation model that outputs 3D characters moving within generated environments rather than video.
The model combines character, motion, and scene generation in one system: users upload a photo to create a character, describe or upload a reference video for motion, and generate the surrounding world from text or a photo.
The model builds the stack in 3D rather than pixel or video space, and generates a 3D scene in seconds.
Instead of generating videos, Viggel’s JST-2 outputs 3D characters that move in their environment.
It combines character, motion, and scene generation into a single model:
Today we are releasing JST-2, the world's first full-stack physics-native world foundation model. Many world models learn in pixel/video space. We went back to first principles and built the entire model stack in 3D. Generate a 3D scene in seconds and expand it in real time. Create characters with motion built in, all through one general architecture.
Source: Santiago · x.comPublished · added here
