Windflow

Hy-Nomas

Tencent Hunyuan · Released 21 October 2026

Hy-Nomas generates a world one video segment at a time and takes free-form direction between segments: add objects, change the style or scene, describe actions or move the camera. A persistent cache keeps the world consistent across turns, and the weights are Apache-2.0.

Try the official demoGet the weightsCan I use it commercially?
Weights
Open weightsApache-2.0
Commercial use
Permissive license
Real-time
No real-time claim
Where to try
1 placeOfficial demo

01 Controls

What you control

  • Free-form text instructions for each segment
  • Camera trajectories
  • Multi-turn changes: new objects, styles, scenes or actions

02 License

Can you use it commercially?

Permissive license

The weights' model card declares Apache-2.0, a permissive license that allows commercial use. Check what the model was built on, because a base model's terms can still apply.

Declared license: Apache-2.0 · Model card

A reading of the published license text, not legal advice.

03 Speed

Does it run in real time?

No real-time claim. Outputs 720×1280 at 24 fps from an 8-step distilled model.

Figures are the authors' own, with their settings and hardware. They are not comparable across models.

04 Setup

How to run it

  1. 01Use the pre-packaged Docker image the README recommends, or install manually with uv sync --extra all.
  2. 02Download tencent/Hy-Nomas from Hugging Face.
  3. 03Run bash scripts/inference/inference_from_config.sh with configs/experiment/inference.yaml; a low-memory config is also provided.
  4. 04Or load tencent/Hy-Nomas-Diffusers with HyNomasModularPipeline from the Diffusers branch the README links.

Summarised from the project's own instructions. Follow the source for exact commands and versions.

05 Access

Where to try it

Listed alphabetically. Windflow maintains this directory and is one of the hosts; a listing does not promise capacity at any given moment.

06 Questions

Hy-Nomas FAQ

Is Hy-Nomas open source?

Hy-Nomas publishes downloadable weights under Apache-2.0. The declared license is permissive.

Can I use Hy-Nomas commercially?

The weights' model card declares Apache-2.0, a permissive license that allows commercial use. Check what the model was built on, because a base model's terms can still apply.

Does Hy-Nomas run in real time?

Outputs 720×1280 at 24 fps from an 8-step distilled model.

Where can I try Hy-Nomas?

You can try it on the official demo.

What does Hy-Nomas take as input?

You give it free-form text instructions for each segment; camera trajectories; and multi-turn changes: new objects, styles, scenes or actions.

How do I run Hy-Nomas?

Following the official instructions: Use the pre-packaged Docker image the README recommends, or install manually with uv sync --extra all. Download tencent/Hy-Nomas from Hugging Face. Run bash scripts/inference/inference_from_config.sh with configs/experiment/inference.yaml; a low-memory config is also provided. Or load tencent/Hy-Nomas-Diffusers with HyNomasModularPipeline from the Diffusers branch the README links.

What are alternatives to Hy-Nomas?

Similar models in this directory include ABot-World (open weights, permissive license), AlayaWorld (open weights, custom terms) and DreamX-World (open weights, permissive license).

Similar world models

All 56 in the directory