LongCat-Video
Summary
LongCat-Video is a foundational 13.6B-parameter video generation model that supports Text-to-Video, Image-to-Video, and Video-Continuation in a single architecture. It emphasizes long-video generation, efficient 720p30fps inference, and strong performance aided by multi-reward RLHF, with Avatar variants (1.5) and a Tech Report. The repository provides installation steps, model downloads, and runnable demos, along with licensing under MIT and ongoing community contributions.