Overview
Viggle JST-2 is a world foundation model released by Viggle AI on October 19, 2026, officially described as the world's first full-stack physics-native world foundation model. Unlike world models that learn in pixel or video space, JST-2 builds its entire model stack in 3D space, can generate 3D scenes in seconds and extend them in real time, and creates characters with built-in motion through a general-purpose architecture. The model is already available via API, and the Viggle platform also offers web and mobile video generation services with character replacement and motion driving. JST-2 is positioned as an underlying model capability for 3D scene and character motion generation, with pricing details subject to the official website.
Key Features
- 3D-native modeling: JST-2 builds its model stack in three-dimensional space, unlike world models that learn in pixel or video space.
- Scene generation in seconds: It can generate 3D scenes in seconds and supports real-time extension, making it suitable for scene construction that requires rapid iteration.
- General-purpose architecture characters: It creates characters with built-in motion through a general-purpose architecture, with characters and motion generated within the same model system.
- API now available: The JST-2 model is now available via API, allowing developers to integrate this capability to build their own applications and workflows.
- Web and mobile: The Viggle platform offers web and mobile video generation services with character replacement and motion driving.
Use Cases
- In the 3D scene concept design phase, use JST-2 to generate scene drafts in seconds and extend the spatial layout in real time.
- When producing character demonstrations with motion, directly generate characters and built-in motion through the general-purpose architecture, reducing the need for separate production steps.
- Developers can integrate JST-2's API into their own toolchains to embed 3D scene generation capabilities into existing workflows.
- Use the Viggle platform on web or mobile for video generation, combining character replacement and motion driving to complete content production.
- When there is a need to quickly validate 3D scene ideas, use the real-time extension capability to adjust the spatial structure while generating.
Pros
- Officially described as the world's first full-stack physics-native world foundation model, with its model stack built in 3D space.
- Scene generation can be completed in seconds and supports real-time extension, suitable for creative workflows that require rapid iteration.
- The general-purpose architecture also covers character creation with built-in motion, with characters and motion generated within the same system.
- The JST-2 model is now available via API, making it easy for developers to integrate into their own applications.
- The Viggle platform offers web and mobile services with character replacement and motion driving, providing multiple access points.
Pricing
JST-2 is accessible through the Viggle API, while the web and mobile services have their own subscription plans. See the official site for current pricing.
Summary
Viggle JST-2 was released on October 19, 2026, and is officially described as the world's first full-stack physics-native world foundation model. Its core features are building its model stack in 3D space, generating 3D scenes in seconds with real-time extension, and creating characters with built-in motion through a general-purpose architecture. The model is now available via API, and the Viggle platform offers web and mobile video generation services with character replacement and motion driving. Pricing details are subject to the official website, and it is suitable for developers and creators interested in 3D scene and character motion generation to learn more.
Version History
- Viggle JST-2 launches: 3D-native world foundation model generating scenes in seconds (2026-10-19): Viggle AI released JST-2 on October 19, 2026, describing it as the first full-stack physics-native world foundation model. Unlike world models that learn in pixel or video space, JST-2 builds its entire stack in 3D, generating 3D scenes in seconds with real-time expansion, and creating characters with built-in motion through a general architecture. JST-2 is available via API, while the Viggle platform continues to offer character replacement and motion-driven video on web and mobile.