Tencent has just released two new 3D models on Hugging Face, and they’re pretty good at generating 3D models. These new additions, Hunyuan3D-DiT-v2-mv and Hunyuan3D-DiT-v2-mini, to the Hunyuan3D family make creating detailed 3D shapes easier than ever before. Whether you’re a game developer, product designer, or just someone who wants to turn pictures into 3D models, these new tools can help you do it quickly and with impressive results.
Table of Contents
Tencent New Hunyuan3D Models
1. Hunyuan3D-DiT-v2-mv
This model lets you create detailed 3D shapes from multiple images. Unlike older AI models that only work with one picture, Hunyuan3D-DiT-v2-mv take input images from different angles: front, back, left, and right and builds a complete 3D model that matches all these views perfectly. This means your 3D models will look right from any angle.
The model is built using a “flow-based diffusion transformer.” It has a really smart brain that understands how 2D images should look in 3D space. At 1.1 billion parameters, it’s powerful enough to capture tiny details while still being practical to run on a good computer.
To get more detailed information about the model and how it works, check out our full article here: Hunyuan3D-DiT-v2-mv.
2. Hunyuan3D-DiT-v2-mini
The other exciting release from Tencent is the Hunyuan3D-DiT-v2-mini model. Don’t let the “mini” fool you; this is still a powerful tool. At 0.6 billion parameters (about half the size of its bigger siblings), it’s designed to run on computers with less power. Key Features of this model are as follows:
The real magic of the mini model is how it uses a 6X latent space compression rate. It needs much less memory (VRAM) to run. If you’ve ever tried using AI tools only to get an “out of memory” error, you’ll appreciate why this matters.
Despite its smaller size, it still creates high-quality 3D models. It’s like having a compact car that still gets you where you need to go, just without all the extra features of a luxury model.
Setting Up the Hunyuan3D-DiT-v2-mini
If you’re interested in the mini version, you can download it directly from Hugging Face. To run it as a user-friendly app on your computer, you can use this command:
python3 gradio_app.py --model_path tencent/Hunyuan3D-2mini --subfolder hunyuan3d-dit-v2-mini --texgen_model_path tencent/Hunyuan3D-2
This will create a web interface where you can upload images and generate 3D models without writing any code.
How These Models Fit into the Hunyuan3D Family
These new releases are part of Tencent’s larger Hunyuan3D 2.0 system. This whole system is designed to help people create amazing 3D content without needing years of training.
The complete 2.0 system has two main parts:
- Shape generation models like Hunyuan3D-DiT-v2-mini or any
- Texture synthesis models that add colors and textures to make things look realistic
Together, they work like a team.
Real-World Uses for Tencent 3D Models
Let’s look at some ways people might use these tools:
1. Game Development
Creating background objects and props for video games traditionally takes a lot of time. With mini model, game developers can quickly generate lots of 3D assets from concept art or reference photos.
2. Product Design
A designer could take pictures of a rough physical prototype from different angles, feed them into Hunyuan3D-DiT-v2-mv, and get a clean 3D model ready for refinement and virtual testing.
3. Education
Students learning about 3D modeling can use these tools to quickly bring their ideas to life without getting bogged down in technical details. They can focus on creativity while the AI handles the complex modeling work.
4. 3D Printing
Someone with a 3D printer could photograph an object they want to duplicate, use Hunyuan3D-DiT-v2-mv to create a 3D model, and then print their own version.
Comparing the Best Hunyuan3D Models
With several models now available, here’s a quick comparison to help you choose the right one:
1. Hunyuan3D-DiT-v2-mv
Best for when you need high-quality models and have pictures from multiple angles. Size: 1.1B
2. Hunyuan3D-DiT-v2-mv-Fast
Best when you need models quickly or are creating many models. Size: 1.1B
3. Hunyuan3D-DiT-v2-mini
Best when you have limited computer resources or want faster generation with still-good quality. Size: 0.6B
4. Hunyuan3D-DiT-v2-0
The original image-to-shape model (not multiview). Size: 1.1B
3D Creation with Hunyuan3D Models
With these new releases, Tencent is pushing the boundaries of what’s possible with AI-generated 3D content. These tools make it possible for anyone to turn pictures into detailed 3D models without specialized training or expensive software. The multiview capabilities of Hunyuan3D-DiT-v2-mv solve one of the biggest challenges in AI-generated 3D: creating models that look good from all angles, not just one.
Meanwhile, the compact size of Hunyuan3D-DiT-v2-mini means more people can use this technology, even without high-end hardware. Whether you’re a professional looking to speed up your workflow or a beginner just starting to explore 3D creation, these new models from Tencent offer exciting possibilities for bringing your ideas into three dimensions.
| Latest From Us
- FantasyTalking: Generating Amazingly Realistic Talking Avatars with AI
- Huawei Ascend 910D Could Crush Nvidia’s H100 – Is This the End of U.S. Chip Dominance?
- Introducing Qwen 3: Alibaba’s Answer to Competition
- Google DeepMind AI Learns New Skills Without Forgetting Old Ones
- Duolingo Embraces AI: Replacing Contractors to Scale Language Learning
