According to Zhitong Finance APP, on April 23rd, one month after the open-source release of the five major models in version 2.0, Tencent Holdings (00700) officially launched the new model of the Hunyuan 3D generation model in version 2.5. The modeling precision has significantly improved, achieving ultra-high-definition geometric detail modeling, with smoother surfaces, sharper edges, and richer details, effectively increasing the geometric resolution to 1024, upgrading from standard definition to high definition.
In terms of model architecture, Hunyuan 3D v2.5 has also undergone a comprehensive upgrade, with the total parameter count increasing from 1 billion to 10 billion, and the number of effective polygons increasing by more than ten times. The Hunyuan 3D AI creation engine has been fully updated to the v2.5 model base, while the free generation quota has doubled to 20 times per day. The Hunyuan 3D generation API has also officially launched on Tencent Cloud, open to enterprises and developers.
The new version of Tencent Hunyuan 3D supports 4K high-definition textures and fine-grained bump mapping, capable of simulating the visual effects of surface undulations of objects. Additionally, Hunyuan 3D has become the first in the industry to achieve multi-view input generation of PBR models, significantly enhancing the quality and realism of the generated outputs.
Furthermore, to meet the needs of professional creators, Hunyuan 3D v2.5 has optimized the skeletal skinning system, supporting automatic skeletal binding and automatic weight assignment under non-standard poses, greatly improving the efficiency of 3D animation generation. The 3D generation workflow features have also been further upgraded, providing professional pipeline templates for text-to-3D and image-to-3D intelligent decimation models, as well as multi-view 3D models.
Tencent Hunyuan actively embraces the open-source ecosystem, with the foundational models of Hunyuan 3D 1.0 and 2.0, as well as accelerated, multi-view, and lightweight models based on the 2.0 model, all open-sourced. The total number of stars on GitHub has exceeded 12,000, continuously enriching the 3D AIGC community

TCTZF

TCEHY

00700

On April 23rd, Tencent Holdings Limited released version 2.5 of the Hunyuan 3D generation model, significantly enhancing modeling precision, supporting ultra-high-definition geometric detail modeling, with an effective geometric resolution of 1024. The total parameter count of the new version increased from 1B to 10B, supporting 4K high-definition textures and fine-grained bump maps, optimizing the skeletal skinning system, and improving 3D animation generation efficiency. At the same time, the Hunyuan 3D generation API has been launched on Tencent Cloud, open to enterprises and developers

- On April 23, Tencent Holdings officially released version 2.5 of its Hongyuan 3D generative model, significantly enhancing modeling precision and achieving ultra-high-definition geometric detail with an effective resolution of 1024.  
- The model architecture has been fully upgraded, increasing the total parameter count from 1 billion to 10 billion and expanding the effective polygon count by over ten times.  
- The new version supports 4K textures, advanced bump mapping, and features like automatic rigging for non-standard poses, while also doubling the free generation quota to 20 times per day for users.

Tencent Holdings' Hongyuan 3D generation model has been upgraded, significantly improving modeling precision