CogVideoX

video

by Zhipu AI · Updated December 1, 2025

Generate with CogVideoX

CogVideoX is Zhipu AI's open-source video generation model producing high-quality video clips with impressive motion quality and temporal coherence. As one of the leading open-source video models, it enables local deployment and fine-tuning, making it popular in research and community workflows.

Best For

Open-source videoLocal deploymentResearchCustom workflowsCommunity fine-tuning

Prompting Tips

  1. 1Fully open-source — can be deployed locally and fine-tuned
  2. 2Start with the 5B model for faster generation, 12B for higher quality
  3. 3Supports both text-to-video and image-to-video
  4. 4Use detailed scene descriptions for better motion quality
  5. 5Great for building custom video generation pipelines

Syntax & Constraints

Natural language prompts. Open-source (Apache 2.0). Available in 2B, 5B, and 12B sizes. Text-to-video and image-to-video. Configurable inference parameters.

Example Prompts

🏔 Landscape & Scenery

Coastal waves at sunset

Ocean waves crashing against a rocky coastline at sunset. Spray and mist catching the golden light. Camera positioned low, watching waves approach and recede. Natural sound implied, peaceful yet powerful atmosphere. 6 seconds.

oceansunsetwaves

Build Prompts for CogVideoX

Related Guides