Wan: Open and Advanced Large-Scale Video Generative Models

2025-03-26

AI Summary

This work addresses the openness, performance, and practicality bottlenecks of large-scale video generation models. Methodologically, we introduce Wan—a family of open-source video foundation models (1.3B/14B parameters)—supporting eight core tasks, including image-to-video generation, instruction-driven editing, and personalized video synthesis. We propose the first efficient video VAE, a scalable diffusion Transformer architecture, and integrate mixed-precision training, dynamic-resolution modeling, and lightweight inference optimization. Additionally, we curate a large-scale, multi-source video dataset and an automated evaluation framework. Experiments demonstrate that Wan-14B outperforms existing open-source and commercial models across multiple internal and external benchmarks. Wan-1.3B achieves deployment on consumer-grade GPUs with only 8.19 GB VRAM. All models, code, and data are fully open-sourced to foster reproducible research and community advancement.

Technology Category

• Computer Vision: Large Vision Models
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Generation

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

This report presents Wan, a comprehensive and open suite of video foundation models designed to push the boundaries of video generation. Built upon the mainstream diffusion transformer paradigm, Wan achieves significant advancements in generative capabilities through a series of innovations, including our novel VAE, scalable pre-training strategies, large-scale data curation, and automated evaluation metrics. These contributions collectively enhance the model's performance and versatility. Specifically, Wan is characterized by four key features: Leading Performance: The 14B model of Wan, trained on a vast dataset comprising billions of images and videos, demonstrates the scaling laws of video generation with respect to both data and model size. It consistently outperforms the existing open-source models as well as state-of-the-art commercial solutions across multiple internal and external benchmarks, demonstrating a clear and significant performance superiority. Comprehensiveness: Wan offers two capable models, i.e., 1.3B and 14B parameters, for efficiency and effectiveness respectively. It also covers multiple downstream applications, including image-to-video, instruction-guided video editing, and personal video generation, encompassing up to eight tasks. Consumer-Grade Efficiency: The 1.3B model demonstrates exceptional resource efficiency, requiring only 8.19 GB VRAM, making it compatible with a wide range of consumer-grade GPUs. Openness: We open-source the entire series of Wan, including source code and all models, with the goal of fostering the growth of the video generation community. This openness seeks to significantly expand the creative possibilities of video production in the industry and provide academia with high-quality video foundation models. All the code and models are available at https://github.com/Wan-Video/Wan2.1.

Problem

• Advancing video generation through scalable diffusion transformer models
• Enhancing performance with novel VAE and large-scale data curation
• Providing open-source models for diverse video production tasks

Innovation

• Novel VAE enhances video generation quality
• Scalable pre-training with large datasets
• Automated metrics for performance evaluation