# Framepack

> Video AI Model by Layer — available on Layer

Highly efficient, open-source I2V model that generates video by predicting the next frame.

Framepack is a highly efficient, open-source Image-to-Video model that generates video by autoregressively predicting the next frame. Developed by the creator of ControlNet, its lightweight design makes it accessible for artists and developers to run on consumer hardware. This model is ideal for the AI art community for its focus on providing a foundational, easily modifiable tool for creating short, coherent animations from a single starting image.

## Specifications

| Property | Value |
|----------|-------|
| Provider | Layer |
| Category | Video |
| Price | 1.3 / megapixel second per megapixel_per_second |
| Aspect Ratios | 16:9, 9:16 |
| Durations | 5s, 10s, 15s, 20s, 25s, 30s, 35s, 40s, 45s, 50s, 55s, 60s |

---

[Try Framepack](https://next.app.layer.ai/me/session/new?baseModelId=FRAMEPACK) | [All Video Models](https://layer.ai/models/video) | [All Models](https://layer.ai/models)