Back to glossary

Term

Hunyuan3D 2.0

Hunyuan3D 2.0 is Tencents open generative 3D model (January 2025). In two stages it produces high-resolution, textured 3D assets from text or image and supports text-, image-, sketch- and portrait-to-3D. The weights are freely available on Hugging Face.

Hunyuan3D 2.0 — explained in more detail

Hunyuan3D 2.0 is a generative 3D model that Tencent released as open source on January 21, 2025. From a single image or a text prompt it produces high-quality 3D assets — including high-resolution texture maps. Its hallmark is a two-stage approach: first an untextured geometry model is generated, then a second stage synthesizes the matching textures. Several modes are supported: text-to-3D, image-to-3D, sketch-to-3D and portrait-to-3D.

The code and pretrained models are openly available on Hugging Face; the Hunyuan3D series has gathered more than three million downloads there since late 2024. The practical payoff, according to Tencent: what used to take days or weeks in 3D production shrinks to minutes.

Example / In practice

Game and product design teams use Hunyuan3D 2.0 to turn a concept image into a first textured 3D mesh — a starting point for prototyping, asset libraries or AR/VR scenes. Because the model is open, it can be integrated into custom pipelines (e.g. ComfyUI or Blender workflows).

Distinction from similar terms

Hunyuan3D produces 3D geometry and textures — unlike image models such as FLUX.1, which output flat 2D images. Among 3D generators it sits alongside Microsofts TRELLIS and commercial services such as Meshy or Rodin.

See everything in one place:3D ModelsOpen-Weight