TripoSR
VASTTripoSR is a feed-forward transformer that reconstructs a textured 3D mesh from a single image in under a second, built on the Large Reconstruction Model design. VAST (Tripo AI) developed it with Stability AI and trained it on a curated subset of Objaverse.
The weights are hosted on Stability AI's Hugging Face account and the code on VAST's GitHub account; the map attributes the model to VAST, whose research group wrote it. TripoSR fills the permissive image-to-3D slot beside TripoSG, its MIT-licensed successor from the same group. Released March 2024; no later TripoSR release exists.
Openness
3 high confidence- weights
- open(model.ckpt downloads from stabilityai/TripoSR on Hugging Face without a gate)
- data
- described(a curated subset of Objaverse, which is itself public under CC-BY, is described
- code
- partial(inference script and a local Gradio app only
- license
- MIT(the code, the weights and the demo, per the README and the model card)
TripoSR's code and weights are MIT-licensed and download freely. Only inference code is published, and the training set is described as a curated Objaverse subset without the subset itself being released.
- https://huggingface.co/api/models/stabilityai/TripoSR?expand[]=downloads&expand[]=likes&expand[]=cardData&expand[]=gated&expand[]=lastModified&expand[]=createdAt&expand[]=siblings recorded 2026-09-27
Hugging Face API record for stabilityai/TripoSR: cardData license mit, gated false, files config.yaml and model.ckpt.
- https://huggingface.co/stabilityai/TripoSR/raw/main/README.md recorded 2026-09-27
Card: "We selected a carefully curated subset of the Objaverse dataset for the training data, which is available under the CC-BY license."; the subset is not published.
- https://ungh.cc/repos/VAST-AI-Research/TripoSR/files/main recorded 2026-09-27
Repository file list: run.py, gradio_app.py and the tsr/ package; no training script or training config.
- https://ungh.cc/repos/VAST-AI-Research/TripoSR/files/main/LICENSE recorded 2026-09-27
LICENSE body: "MIT License / Copyright (c) 2024 Tripo AI & Stability AI".
- https://ungh.cc/repos/VAST-AI-Research/TripoSR/files/main/README.md recorded 2026-09-27
README: "The model is released under the MIT license, which includes the source code, pretrained models, and an interactive online demo."; only inference and a Gradio app are documented.
Adoption
3 high confidenceAdoption is measured as Hugging Face downloads of the stabilityai/TripoSR checkpoint, the only first-party copy; third-party conversions are left out.
- https://huggingface.co/api/models/stabilityai/TripoSR?expand[]=downloads&expand[]=likes&expand[]=cardData&expand[]=gated&expand[]=lastModified&expand[]=createdAt&expand[]=siblings recorded 2026-09-27
Hugging Face API record for stabilityai/TripoSR: downloads 201059.
Capability
2 medium confidenceTripoSR places in the lower half of the image-to-3D arena, below Hunyuan3D-2 and 2.1 in its upper half and well below TRELLIS near the top. It trades quality for speed, reconstructing a mesh in under a second.
- https://3d-arena-3d-arena.hf.space/api/leaderboard?vote_type=render recorded 2026-09-27
Render board of 24 entries: "TripoSR" rank 20, score 1081, 11388 votes; "Hunyuan3D-2" rank 9, "Hunyuan3D-2.1" rank 10.
- https://3d-arena-3d-arena.hf.space/api/leaderboard?vote_type=topology recorded 2026-09-27
Topology board of 21 entries: "TripoSR" rank 20, score 981.
Verified 2026-09-27