---
title: Zing-0.5
canonical_url: "https://www.modelscope.cn/models/seedleap/Zing-0.5"
md_url: "https://www.modelscope.cn/models/seedleap/Zing-0.5.md"
repository: seedleap/Zing-0.5
last_updated: 2026-09-17
license: apache-2.0
pipeline_tag: text-to-video-synthesis
tasks:
  - text-to-video-synthesis
base_model:
  - Wan-AI/Wan2.2-TI2V-5B-Diffusers
base_model_relation: finetune
parameters: 6.4B
tensor_type:
  - F32
  - BF16
library_name:
  - safetensors
  - diffusers
downloads: 1794
stars: 62
tags:
  - world-model
  - video-generation
  - text-to-video
  - image-to-video
  - action-conditioned
  - causal
---

# Zing-0.5

> Zing-0.5 - seedleap 在 ModelScope 开源的模型。Zing-0.5: Toward Playable Worlds with Real-Time Joint Action and Text Control

seedleap/Zing-0.5 是 ModelScope 魔搭社区上的 6.4B 参数text-to-video-synthesis模型，采用 apache-2.0 许可，基于 Wan-AI/Wan2.2-TI2V-5B-Diffusers 构建。

- **Repository**: seedleap/Zing-0.5
- **License**: apache-2.0
- **Tasks**: text-to-video-synthesis
- **Parameters**: 6.4B
- **Base model**: Wan-AI/Wan2.2-TI2V-5B-Diffusers
- **Tags**: world-model, video-generation, text-to-video, image-to-video, action-conditioned, causal
- **Downloads**: 1794
- **Stars**: 62
- **Last updated**: 2026-09-17

Source: https://www.modelscope.cn/models/seedleap/Zing-0.5

---

# Zing-0.5: Toward Playable Worlds with Real-Time Joint Action and Text Control

[Project Page](https://zing.loopit.me/) · [GitHub](https://github.com/seedleap/zing-world-model) · [Tech report](https://arxiv.org/abs/2609.17909) · [SGLang Inference](https://github.com/seedleap/Zing-SGLang)

Zing-0.5 is a 5B causal world model developed by the Seedleap.ai team (涌跃智能) for real-time interaction. It continuously predicts the visual world from its current state while text prompts and keyboard actions alter the scene, motion, and future evolution during generation.

Zing-0.5 supports text-initialized generation, single-image initialization, prompt changes during a rollout, and continuous W/A/S/D/I/J/K/L keyboard control. Causal KV caching and four-step DMD sampling enable responsive long-horizon generation on a single GPU.

## Model Files

Download the model repository with the following directory structure:

```text
Zing-0.5/
├── generator/
│   └── model.pt
└── pretrained/
    ├── text_encoder/
    ├── tokenizer/
    └── vae/
```

`generator/model.pt` must directly contain the generator state dict. Parameter names and shapes are loaded strictly.

## Inference

For realtime serving through SGLang's WebSocket API, use
[Zing-SGLang](https://github.com/seedleap/Zing-SGLang).

Use the standalone inference code from the [Zing GitHub repository](https://github.com/seedleap/zing-world-model):

```bash
git clone https://github.com/seedleap/zing-world-model.git
cd zing-world-model

ZING_MODEL=/path/to/Zing-0.5
CUDA_VISIBLE_DEVICES=0 \
ZING_PYTHON=/path/to/python \
bash run.sh \
  --pretrained-dir "$ZING_MODEL/pretrained" \
  --checkpoint "$ZING_MODEL/generator/model.pt" \
  --messages examples/case3_action_t2v.jsonl \
  --output-dir outputs/case3 \
  --seed 0
```

The code repository includes ready-to-run Action T2V and Action TI2V JSONL examples with the required reference images.

## GPU Memory

| GPU memory | `local_attn_size` | `sink_size` |
| --- | ---: | ---: |
| 80 GB or more | `97` | `9` |
| Less than 80 GB | `33` | `5` |

The default `97/9` configuration has been validated on a single NVIDIA H100 80 GB. Use `33/5` on GPUs with less memory. Full-history attention is available with `--local-attn-size -1 --sink-size 0`.

## Output

Each JSONL row produces one result named from `sample_id`. Multi-frame rollouts are saved as H.264 MP4 at 24 FPS.

## Limitations

Long rollouts may exhibit visual drift or physical inconsistencies. Action responsiveness can vary with scene content and viewpoint.

## License

Zing-0.5 is released under the [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0).