Free & Open Source • 13.6B Parameters • MIT Licensed
Unified architecture supporting Text-to-Video, Image-to-Video, and Video-Continuation within a single framework. No need for multiple models.
Generate 720p 30fps videos within minutes using coarse-to-fine strategy and Block Sparse Attention for high-resolution efficiency.
Natively pretrained on Video-Continuation tasks, enabling minutes-long videos without color drifting or quality degradation.
Powered by Group Relative Policy Optimization (GRPO), achieving performance comparable to leading commercial solutions.
Completely free and open source. Use commercially, modify freely, and deploy anywhere without restrictions.
13.6B dense parameters outperforming 28B MoE models. All parameters are activated, ensuring consistent quality.
| Model | Accessibility | Architecture | Total Params | Text-Alignment ↑ | Visual Quality ↑ | Motion Quality ↑ | Overall Quality ↑ |
|---|---|---|---|---|---|---|---|
| Veo3 | Proprietary | - | - | 3.99 | 3.23 | 3.86 | 3.48 |
| PixVerse-V5 | Proprietary | - | - | 3.81 | 3.13 | 3.81 | 3.36 |
| Wan 2.2-T2V-A14B | Open Source | MoE | 28B (14B activated) | 3.70 | 3.26 | 3.78 | 3.35 |
| LongCat-Video | Open Source | Dense | 13.6B | 3.76 | 3.25 | 3.74 | 3.38 |
| Model | Accessibility | Architecture | Total Params | Image-Alignment ↑ | Text-Alignment ↑ | Visual Quality ↑ | Motion Quality ↑ | Overall Quality ↑ |
|---|---|---|---|---|---|---|---|---|
| Seedance 1.0 | Proprietary | - | - | 4.12 | 3.70 | 3.22 | 3.77 | 3.35 |
| Hailuo-02 | Proprietary | - | - | 4.18 | 3.85 | 3.18 | 3.80 | 3.27 |
| Wan 2.2-I2V-A14B | Open Source | MoE | 28B (14B activated) | 4.18 | 3.33 | 3.23 | 3.79 | 3.26 |
| LongCat-Video | Open Source | Dense | 13.6B | 4.04 | 3.49 | 3.27 | 3.59 | 3.17 |
git clone https://github.com/meituan-longcat/LongCat-Video
cd LongCat-Video# Create conda environment
conda create -n longcat-video python=3.10
conda activate longcat-video
# Install torch
pip install torch==2.6.0+cu124 torchvision==0.21.0+cu124 --index-url https://download.pytorch.org/whl/cu124
# Install flash-attn-2
pip install ninja psutil packaging flash_attn==2.7.4.post1
# Install other requirements
pip install -r requirements.txtpip install "huggingface_hub[cli]"
huggingface-cli download meituan-longcat/LongCat-Video --local-dir ./weights/LongCat-Video# Single-GPU inference
torchrun run_demo_text_to_video.py --checkpoint_dir=./weights/LongCat-Video --enable_compile
# Multi-GPU inference
torchrun --nproc_per_node=2 run_demo_text_to_video.py --context_parallel_size=2 --checkpoint_dir=./weights/LongCat-Video --enable_compileCreate engaging videos for Instagram, TikTok, and YouTube from simple text prompts.
Generate educational videos and visual explanations for online courses and tutorials.
Produce product demos and promotional content without expensive video production.
Bring your artistic visions to life with AI-powered video generation.
Experiment with video generation models and advance AI research.
Scale video content production for media companies and agencies.