Download
LongCat-Video is a 13.6-billion-parameter foundation model for generating and extending video. A unified architecture handles text-to-video, image-to-video, and video-continuation tasks without separate models. It is pretrained for continuation, allowing it to create minutes-long sequences while limiting color drift and quality loss. A coarse-to-fine strategy operates across time and space to produce 720p video at 30 frames per second efficiently. Block Sparse Attention reduces high-resolution inference costs, while multi-reward GRPO training improves visual quality and prompt alignment. The repository includes inference scripts for single- and multi-GPU execution, model-download instructions, and interactive generation examples. It also provides audio-driven Avatar variants for expressive single- or multi-character animation.
Features
- Unified text-to-video, image-to-video, and continuation generation
- Minutes-long video generation with visual consistency
- Efficient 720p output at 30 frames per second
- Block Sparse Attention for high-resolution inference
- Single-GPU and multi-GPU execution modes
- Audio-driven single- and multi-character animation
Project Samples
