# Wan 3.0: Leading AI Video Generation Model

[![Official Website](https://img.shields.io/badge/Official_Website-wan.video-blue)](https://wan.video/)
[![Wan Explore](https://img.shields.io/badge/Explore-wan.video-blue)](https://create.wan.video/explore)
[![QwenCloud Models](https://img.shields.io/badge/QwenCloud-wan_models-orange)](https://www.qwencloud.com/models?search=wan)

[Wan 3.0](https://wan.video/) is a state-of-the-art AI video generation model powered by Alibaba Cloud. It pushes the boundaries of open-source video synthesis by delivering cinematic realism, advanced editing capabilities, and comprehensive multi-modal generation tools.

---

## 🌟 Key Features

* **Native 30s Duration**: Supports seamless 30-second video generation with intelligent duration control, enabling full storytelling and creative freedom.
* **Omni-Creation**: Generates video content from up to 20 reference assets, including complex documents, webpage parsing, text, and images.
* **Cinematic Realism & Sound**: Enhances visual texture, physical dynamics, and integrated sound design for immersive audiovisual quality.
* **Pixel-Perfect Consistency**: Accurately preserves character and reference details across frames to maintain production-grade certainty.
* **Precision Video Editing**: Offers instruction-based and reference-based video editing, supporting multi-person/human-object interactions and auto scene splitting.
* **Portrait Customization**: Fine-grained control over bone structure, eye details, and subtle facial expressions.
* **Sequential Storytelling & Interactive Editing**: Generates up to 12 sequential images with cohesive styling, alongside pixel-level box-selection editing.
* **Advanced Text Rendering**: Renders long-form text across 12 languages, including support for complex charts, formulas, and infographics.

---

## 🚀 Quick Start & Modes

Wan 3.0 offers multiple creative pipelines:

| Mode | Description |
| :--- | :--- |
| **Text to Video** | Cinematic precision driven by complex prompt execution. |
| **Image to Video** | Converts static images to 30s clips with auto scene splitting. |
| **Reference to Video** | Casts multi-character references into new dynamic scenes. |
| **Video Editing** | Re-creates visuals, storylines, and environmental elements based on prompts. |
| **Text to Image / Image Edit** | High-fidelity photorealistic generation and conversational image editing. |

* Explore creations and showcase on **[Wan Explore](https://create.wan.video/explore)**.

---

https://github.com/user-attachments/assets/9fa896c8-3ed5-4df4-b014-c092ef0bbad2




https://github.com/user-attachments/assets/d97ffc4c-bb2d-4685-b422-a2e42c890104




https://github.com/user-attachments/assets/f0283668-3446-4c19-b959-cb96181d1f49




https://github.com/user-attachments/assets/6aa06886-21e3-48db-ae40-02d31007686f




* **Powered by**: Alibaba Cloud
* **Policies**: Please review our [Terms of Service](https://wan.video/) and [Privacy Policy](https://wan.video/) before usage.