TRELLIS.2-4B is Microsoft's latest model for native and compact structured latents for 3D generation. It converts images into 3D representations efficiently. The model offers improved inference speed and quality compared to previous approaches and is designed for practical 3D content creation workflows. It is available on Hugging Face with supporting code in the official GitHub repository.
In the Other AI space, TRELLIS.2 4B takes a focused approach. It focuses on generating high-quality 3D assets from single images using compact structured latent representations. It is built as an open-source project for 3D content creators and AI researchers. The project is open source (MIT). It ships for the web, the command line, and API.
Microsoft builds and maintains TRELLIS.2 4B, and it first shipped in 2025. Development happens publicly on GitHub with 8.8k stars and 4 commits in the last 90 days. PulseGate's similarity index places it among 7 comparable projects. Among its 3 catalogued features are image-to-3D, Structured Latents, and 3D Generation.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do