#4141st of 70 in Open modelsFor image, video and design, 3d
hunyuan3d-1
Text-to-3D and image-to-3D mesh generation from multi-view diffusion
Live demo ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 3.5k
- License
- custom license
- Last commit
- Nov 2026
- Language
- Python
3,438 on 21 Apr 20263,507 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Hunyuan3D-1.0 generates textured 3D meshes from a text prompt or a single image. It runs a multi-view diffusion model, then a sparse-view reconstruction model, with Hunyuan-DiT handling text-to-image. Lite and standard variants ship with a CLI and a Gradio demo; the README cites about 10 s (lite) and 25 s (standard) per mesh.
Who it is for: Engineers with a 16 GB+ NVIDIA GPU trying open text/image-to-3D models
Strengths
- Accepts text or image input, with bilingual Chinese and English prompts
- Lite and standard variants; split-module scripts fit 10-16 GB GPUs
- Includes CLI, Gradio demo, texture mapping and optional texture baking
- Checkpoints on Hugging Face plus a hosted demo space
Weaknesses
- Needs 30 GB VRAM for standard, 22 GB for lite; --save_memory slows inference
- Heavy install: CUDA 12.1 PyTorch, pytorch3d, nvdiffrast and a dust3r checkout
- Baking depends on DUSt3R under CC BY-NC-SA 4.0, so non-commercial only
- Gradio demo has no auth; pinned torch 2.6.0 has a known CVE
What it needs
- GPU required
- Needs PyTorch (CUDA 12.1), pytorch3d, nvdiffrast, DUSt3R, Hunyuan-DiT weights, huggingface-cli
- Models: Hunyuan3D-1 lite, Hunyuan3D-1 std, Hunyuan3D-1 svrm, HunyuanDiT-v1.1-Diffusers-Distilled
- port 8080
Quick answers
Is hunyuan3d-1 open source?
Its code is public on GitHub under its own license terms; read them before you use it at work.
Does hunyuan3d-1 need a GPU?
Yes, it needs a GPU.
Also in Open models
See all 70| Rank | Project | Score |
|---|---|---|
| 1 | ruviewWiFi CSI sensing on ESP32 for presence, breathing, heart rate and pose | 94 out of 100 |
| 2 | layaSingle-pass multilingual classifier for typed choice, score and yes/no decisions | 88 out of 100 |
| 3 | rf-detrReal-time transformer models for object detection, segmentation and keypoints | 79 out of 100 |
| #4 | timesfmPretrained time-series foundation model for zero-shot forecasting from Google Research | 78 out of 100 |
| #5 | yolov5PyTorch YOLOv5 object detection, segmentation and classification models with training code | 76 out of 100 |