FIFoundationIndex
IndexMapAbout
Company / HubVideoImage

Runway

Also known as: RunwayML · Gen-3 Alpha · Gen-4

Plain English

One of the earliest and most influential AI video labs—co-creator of the original Stable Diffusion research and builder of the Gen-2, Gen-3, and Gen-4 video generation models geared toward filmmakers and studios.

Technical Overview

Runway co-authored the foundational Latent Diffusion Models paper before focusing on multimodal Video Generation Models (Sora, Veo, Kling, Wan) and world-simulation models. Its Gen-3 and Gen-4 systems pair spatiotemporal DiT backbones with fine-grained camera motion, Act-One performance transfer, and reference-consistent character/scene control.

References:Runway Research

Chronological Entries (2)

#1Release
Curated

Runway launched Gen-3 Alpha, a major leap in temporal consistency and expressive human motion trained jointly on video and image data, followed by Turbo and Act-One character performance controls for production filmmakers.