Overview
- Alibaba has made its Wan 2.1 text-to-video AI models open-source, allowing free access for researchers, developers, and commercial entities worldwide.
- The Wan 2.1 series includes four models, with the most advanced supporting 14 billion parameters and ranking as the top video-generation model on the VBench Leaderboard.
- The models can generate realistic videos with complex motion, cinematic quality, and sound effects, and support both Chinese and English text in AI-generated videos.
- Smaller versions of the models are optimized for consumer-grade hardware, enabling broader accessibility without requiring powerful GPUs.
- Concerns have been raised about the potential misuse of such tools, as Alibaba has not outlined safety measures or methods to identify AI-generated content.