L
LongCat-Video
Description
LongCat-Video is an open video generation model from Meituan (the LongCat team) with 13.6 billion parameters. A single model handles three tasks: text-to-video, image-to-video and continuing an existing clip. The weights are published on Hugging Face under the MIT license, which allows commercial use. Sound is not claimed in the author's model card. The author does not state VRAM requirements, and the model card has no official demo of the base model without installation.
Pros and cons
Pros
- MIT license, commercial use allowed
- Three tasks in one model: text-to-video, image-to-video, video continuation
- Weights are not behind an access gate
Cons
- The author does not state VRAM requirements
- No official demo without installation
- Sound is not claimed in the author's model card
- Released in October 2025, no newer versions
Key facts
- License
- MIT
- Commercial use
- Allowed
- Developer
- Meituan
- Interface languages
- English
- Checked on
- Oct 17, 2026
Editorial score
How strongly the AI Expertise editors recommend this open model for each of its tasks right now, on a 0–10 scale. The score combines output quality, capabilities, ease of launch, license openness and freshness.