From 137432e82a1cecf215b9283018178d3237a154a0 Mon Sep 17 00:00:00 2001
From: yycoco <1428275037@qq.com>
Date: Thu, 26 Mar 2026 10:58:23 +0800
Subject: [PATCH] Add MagicWorld to Community Works

---
 README.md | 1 +
 1 file changed, 1 insertion(+)

diff --git a/README.md b/README.md
index 3aa29f54..d36cfdd8 100644
--- a/README.md
+++ b/README.md
@@ -47,6 +47,7 @@ We are excited to introduce **Wan2.2**, a major upgrade to our foundational vide
 ## Community Works
 If your research or project builds upon [**Wan2.1**](https://github.com/Wan-Video/Wan2.1) or [**Wan2.2**](https://github.com/Wan-Video/Wan2.2), and you would like more people to see it, please inform us.
 
+- [MagicWorld](https://github.com/vivoCameraResearch/Magic-World), built on **Wan2.1**, is an autoregressive interactive video world model designed to address motion drift and long-horizon error accumulation. It enables more realistic dynamics and stable scene evolution on the RealWM120K dataset. MagicWorld achieves 18 FPS on an L40s GPU and delivers superior performance on VBench (RealWM120K-Val). Visit the project [webpage](https://vivocameraresearch.github.io/magicworld/) for more details.
 - [Prompt Relay](https://github.com/GordonChen19/Prompt-Relay), a plug-and-play, inference-time method for temporal control in video generation. Prompt Relay improves video quality and gives users precise control over what happens at each moment in the video. Visit their [webpage](https://gordonchen19.github.io/Prompt-Relay/) for more details.
 - [Helios](https://github.com/PKU-YuanGroup/Helios), a breakthrough video generation model base on **Wan2.1** that achieves minute-scale, high-quality video synthesis at 19.5 FPS on a single H100 GPU (about 10 FPS on a single Ascend NPU) —without relying on conventional long video anti-drifting strategies or standard video acceleration techniques. Visit their [webpage](https://pku-yuangroup.github.io/Helios-Page/) for more details.
 - [LightX2V](https://github.com/ModelTC/LightX2V), a lightweight and efficient video generation framework that integrates **Wan2.1** and **Wan2.2**, supporting multiple engineering acceleration techniques for fast inference. [LightX2V-HuggingFace](https://huggingface.co/lightx2v), offers a variety of Wan-based step-distillation models, quantized models, and lightweight VAE models.