Key Features

Provides an open multimodal world-model framework.
Supports reconstructing, generating, and simulating 3D worlds.
Targets reusable 3D scene representations.
Useful for games, robotics, and virtual production.
Combines multimodal input handling with 3D world modeling.
Provides public source code through GitHub.
Can support digital twin and embodied AI workflows.
Addresses multiple stages of 3D world creation.

The system combines multimodal inputs with 3D world modeling, allowing it to reconstruct existing environments, generate new ones, and support simulation workflows. Technically, this requires geometry representation, scene semantics, temporal or interactive state, and generation mechanisms that preserve world consistency. As a GitHub release, it provides a basis for developers to inspect and adapt the implementation.


HY World 2.0 is valuable because it targets the full lifecycle of 3D worlds rather than only one step such as image generation or reconstruction. It can support research into reusable worlds that can be viewed, simulated, and modified.

Get more likes & reach the top of search results by adding this button on your site!

Embed button preview - Light theme
Embed button preview - Dark theme
TurboType Banner
Zero to AI Engineer Program

Zero to AI Engineer

Skip the degree. Learn real-world AI skills used by AI researchers and engineers. Get certified in 8 weeks or less. No experience required.

Subscribe to the AI Search Newsletter

Get top updates in AI to your inbox every weekend. It's free!