50+
Papers at top venues
4
Papers in 2026
30+
Open-source projects
30K+
GitHub stars
Day-one
Paper · weights · code
What we are working on.
World model initiative
Research preview of an interactive world model. Spaces that keep running after you leave them.
Read more
SIGGRAPH 2026
Open-source 3D Gaussian generator. You set the quality budget, and there is almost nothing to install.
GitHub
Open source
Open model, image to mesh. The one most of the ecosystem builds on.
GitHub
Open source · Stability AI
Textured meshes without a GPU. Built with Stability AI; 15,000 stars, developers in 157 countries.
GitHub
From the community
UniRig took a week of weight painting out of every character we ship. We have not opened the old tool since.
Kai Brennan, Technical Director — Longhand
Rigs roughly two hundred characters a year on an open-source release
Published, then released.
Four papers this year. Weights, code and plugins go out on the day of the paper.
May 8, 2026
Runjie Yan, Yan-Pei Cao, Peng Wang, Ding Liang, Yuan-Chen Guo
Introduces Density-Sampled Gaussians for adaptive, grid-free 3D generation. The VecSeq diffusion framework resolves permutation ambiguity by anchoring latents to a deterministic 3D Sobol sequence, which enables variable-resolution decoding from a single compact latent code.
April 9, 2026
Yi-Hua Huang, Zi-Xin Zou, Yuting He, Chirui Chang, Cheng-Feng Pu, Ziyi Yang, Yuan-Chen Guo, Yan-Pei Cao, Xiaojuan Qi
Produces animatable 3D assets directly from a single image, skipping the generate-then-rig pipeline. Shape, skeleton and skinning weights unify into a continuous S³ Fields representation, and the Dual Skin Field handles rigs of arbitrary complexity.
March 2, 2026
Hanxiao Wang, Yuan-Chen Guo, Ying-Tian Liu, Zi-Xin Zou, Biao Zhang, Weize Quan, Ding Liang, Yan-Pei Cao, Dong-Ming Yan
Addresses the quadratic compute bottleneck by moving to a higher semantic tier. A one-face-one-token strategy cuts sequence length ninefold, reaching a compression ratio of 0.11, and couples an autoregressive autoencoder with latent diffusion.
February 4, 2026
Jia-Peng Zhang, Cheng-Feng Pu, Meng-Hao Guo, Yan-Pei Cao, Shi-Min Hu
Compresses sparse skinning matrices into discrete token sequences. The TokenRig framework synthesizes skeleton topology and surface deformation as one sequence, post-trained with reinforcement learning, and reports up to 133% improvement in skinning accuracy.
News this year.
June 2026
First public preview of the world model: persistent, multi-agent, authorable from any input.
Read
May 2026
Single image to 3D, with weights and a ComfyUI node out on the same day as the paper.
Read
March 2026
UniRig becomes the global benchmark for automatic rigging across arbitrary character topologies.
Read
Open research roles.
Dedicated compute, competitive pay, and papers published as they land.
Research Scientist — World Models
Beijing / ShanghaiApply
Research Scientist — 3D Generation
Beijing / Shanghai / HangzhouApply
Senior Research Scientist — 3D Vision & Graphics
Beijing / Shanghai / HangzhouApply
Research Engineer — Inference Acceleration
BeijingApply
Research Intern
From ¥1,000 a day · from ¥2,000 for top-venue first authors
Beijing / Shanghai / HangzhouApply
