Industry's first, Tencent's hybrid 3D world model officially released and open-sourced

July 27, 2012 - Today, at 2025 World Artificial Intelligence ConferenceTencentAt the forum, Tencent officially released the hybrid 3D World Model 1.0 and fullyOpen Source.

According to the introduction,This is the industry's first open-source immersive roaming, interactive, and simulatable world generation model, bringing new possibilities to game development, VR, digital content creation and other fields. In addition to the 3D world model, Tencent Mixed Element also disclosed a series of open source plans including end-side hybrid reasoning language models, multimodal understanding models, game vision models, and more.

Industry's first, Tencent's hybrid 3D world model officially released and open-sourced

Tencent Hybrid 3D World Model 1.0 integrates panoramic visualization and hierarchical 3D reconstruction technologies, and supports both text and image input, realizing the generation of high-quality and diverse styles of roamable 3D scenes.3D virtual worlds that used to take a team of professional modelers weeks to build can now be created in minutes with a single word or image!.

For game developers, the hybrid 3D world model greatly simplifies the process of building 3D scenes. With simple commands, the model can quickly generate a complete 3D scene with buildings, terrain, and vegetation. The output Mesh file can be used for game prototyping or level design, and can be flexibly adjusted for foreground objects and sky backgrounds to meet the needs of personalized creation.

With the Hybrid 3D creation engine, ordinary users with no modeling experience can quickly generate a 360° immersive visual space in just one sentence or one picture, and the generated scene can be seamlessly imported into a virtual headset such as Vision Pro to bring an immersive experience.

1 From the official introduction, we learned that the core of the Hybrid 3D World Model 1.0 also lies in its innovative "Semantic Hierarchical 3D Scene Representation and Generation Algorithm". The algorithm deconstructs the complex 3D world into different semantic layers.Achieve intelligent separation of foreground and background, ground and skyAIGC's 3D mesh assets are not only visually realistic, but also output standardized 3D mesh assets that are compatible with Unity, Unreal Engine, Blender, and other mainstream tools. Users can independently edit or physically simulate elements within the scene, seamlessly connecting AIGC technology with traditional CG workflows.

Compared to the world's leading open source models, the Hybrid 3D World Model 1.0 is aesthetically superior in terms of aesthetics and command adherence in both the textual and graphical worlds.Comprehensively surpasses current SOTA open source models.

Tencent Hybrid 3D World Model 1.0:

  • Project homepage: https://3d-models.hunyuan.tencent.com/world/
  • Experience: https://3d.hunyuan.tencent.com/sceneTo3D
  • Hugging Face model address: https://huggingface.co/tencent/HunyuanWorld-1
  • Github project address: https://github.com/Tencent-Hunyuan/HunyuanWorld-1.0
statement:The content of the source of public various media platforms, if the inclusion of the content violates your rights and interests, please contact the mailbox, this site will be the first time to deal with.
Information

Ali's first self-developed Quark AI glasses unveiled, built-in Alipay "look and pay"

2025-7-27 13:32:47

HeadlinesInformation

The Chinese government has initiated the establishment of the World Artificial Intelligence Cooperation Organization (WACICO), with initial consideration of a headquarters in Shanghai.

2025-7-27 13:37:02

Search