Stable Video Diffusion

Stability AI's open-source latent diffusion model transforms static images into high-resolution cinematic clips with precise motion control.

Stability AI released Stable Video Diffusion (SVD) as a foundational architecture for high-quality video synthesis. The suite features two primary checkpoints: SVD (14 frames) and SVD-XT (25 frames). These models generate video at resolutions up to 576x1024 with adjustable frame rates between 3 and 30 fps. SVD integrates directly with the Stable Diffusion ecosystem to facilitate image-to-video workflows and camera motion conditioning. Developers access the weights through Hugging Face to build custom applications for research or commercial production.

https://stability.ai/news/stable-video-diffusion-open-ai-video-model