- Home
- Text-to-Video Generators
- CogVideoX
On the recordRead 13 Oct 202627 facts
CogVideoX
On the record
- Commercial use
- The CogVideoX License requires commercial users to register for a basic commercial license and limits service visits to one million per month under that license.github.com · read 13 Oct 2026
- Commercial use limit
- The model license permits registered basic-license commercial activities up to 1 million service visits per month; higher usage requires contacting the business team for more licenses.github.com · read 12 Oct 2026
- Deployment
- The repository provides inference and fine-tuning code through SAT and Diffusers, plus a simple Gradio web application.github.com · read 12 Oct 2026
- Free Colab examples
- The repository provides Colab examples described as runnable on free T4 instances, including text-to-video and image-to-video notebooks.github.com · read 13 Oct 2026
- Generation tasks
- The project lists text-to-video generation, video continuation, and image-to-video generation as supported tasks.github.com · read 13 Oct 2026
- Hardware limits
- The README lists minimum Diffusers memory usage of 4 GB for CogVideoX1.5-5B-I2V and 5 GB for CogVideoX-5B, with quantized configurations listed lower.github.com · read 13 Oct 2026
- Hardware requirements
- The repository lists minimum Diffusers memory figures as low as 4 GB for CogVideoX-2B and notes that INT8 can reduce memory needs while slowing inference.github.com · read 12 Oct 2026
- Image-to-video
- Yesgithub.com · read 19 Sep 2026
- Integrations
- The project lists community integrations including ComfyUI-CogVideoXWrapper, CogStudio, and Xorbits Inference.github.com · read 13 Oct 2026
- License
- The code and CogVideoX-2B are released under Apache 2.0, while CogVideoX-5B model weights use the CogVideoX License.github.com · read 13 Oct 2026
- License restrictions
- The model license prohibits use for military or illegal purposes and certain acts described in the license, and provides the software without warranty.github.com · read 12 Oct 2026
- Licensing
- The code is released under Apache 2.0, while model weights have a separate license that allows academic research use and requires a basic commercial license for commercial use.github.com · read 12 Oct 2026
- Maker
- Zhipu’s about page lists its founding year as 2019 and its address in Haidian District, Beijing, China.zhipuai.cn · read 13 Oct 2026
- Maximum video length
- 0.1666666667github.com · read 19 Sep 2026
- Models
- The repository lists CogVideoX1.5-5B, CogVideoX1.5-5B-I2V, CogVideoX-2B, CogVideoX-5B, and CogVideoX-5B-I2V.github.com · read 13 Oct 2026
- Online access
- The repository links online CogVideoX-5B demos on Hugging Face Spaces and ModelScope Spaces.github.com · read 12 Oct 2026
- Online demo
- The project links to Hugging Face and ModelScope Spaces to try CogVideoX-5B online.github.com · read 13 Oct 2026
- Prompt language
- The model supports English input, and the README says other languages can be translated into English for use.github.com · read 13 Oct 2026
