Runway unveils Gen-3 Alpha: advanced AI model for hyper-realistic 10-second video clips

Runway has introduced Gen-3 Alpha, its latest AI video model capable of producing hyper-realistic 10-second clips. This model represents a significant advancement over its predecessor, Gen-2, designed to deliver high-fidelity and controllable video outputs, enhance AI video generation technology.

The Gen-3 Alpha model is the first to be trained on a new infrastructure tailored for large-scale multimodal training, which improves the fidelity, consistency, and motion of the generated videos. It allows users to customize videos for specific styles and consistent characters, excelling in generating expressive human characters with a wide range of actions and emotions. This new model directly competes with other highly talked-about alternatives from last week, such as the new Dream Machine from Luma AI.

The new model will be integrated into various Runway tools such as text-to-video, image-to-video, text-to-image, Motion Brush, Advanced Camera Controls, and Director Mode. Developed using curated datasets by Runway’s research team, the model includes new safeguards adhering to Content Credentials (C2PA) provenance standards. The model will be available to Runway subscribers, Creative Partners Program members, and Enterprise users this week.