Tencent releases and open-sources the Hunyuan 3D World Model 2.0
On April 16th, Tencent released and open-sourced the Hunyuan 3D World Model 2.0 (HY-World 2.0). This model is a multimodal world model that supports generating 3D worlds based on inputs such as text, images, and videos, and can export 3D assets in multiple formats, suitable for the rapid generation of game maps and level prototypes. Compared to the previous model, HY-World 2.0 can generate editable 3D asset files, enhancing its practicality, and is now available on the official Tencent Hunyuan 3D Creation Engine website, where users can apply for access
According to Zhitong Finance APP, on April 16th, Tencent (00700) officially released and open-sourced the HY-World 2.0, a mixed yuan 3D world model. HY-World 2.0 is a multimodal world model that can automatically generate, reconstruct, and simulate 3D worlds based on different types of inputs such as text, images, and videos. It also supports the export of multi-format 3D assets (Mesh/3DGS/point clouds, etc.) and seamlessly integrates with existing game workflows for rapid generation of game maps and level prototypes. The mixed yuan world model 2.0 allows AI to evolve from "dialogue," "drawing," and "shaping objects" to "creating worlds"—generating complete "worlds" that include people, objects, and scenery.
Previously, mainstream world models, such as Google's Genie 3 and the mixed yuan world model 1.5, could only generate video files. The practicality of the newly released world model 2.0 has advanced further—it can directly generate editable 3D asset files, and these generated 3D assets can be directly imported into game production or embodied simulation engines, achieving true playability and usability.
Currently, the world model 2.0 has been launched on the Tencent mixed yuan 3D creation engine official website, and users can apply for access
