Runway Unveils Gen-3 Alpha to Advance AI Video Generation
Runway launches Gen-3 Alpha, a new foundation model that delivers major improvements in video fidelity, consistency, and motion. The system powers various creative tools and aims to pave the way for General World Models.
Runway introduces Gen-3 Alpha as the first of a next generation of foundation models trained on a new infrastructure built for large-scale multimodal training. This new model represents a major improvement in fidelity, consistency, and motion over the previous Gen-2 system, bringing the company closer to its goal of building General World Models. All the demonstration videos showcase the model's raw output without any modifications.
The system trains jointly on videos and images to power Runway's Text to Video, Image to Video, and Text to Image tools. It also supports existing control modes like Motion Brush, Advanced Camera Controls, and Director Mode, while preparing to launch upcoming tools for fine-grained control over structure, style, and motion. Additionally, Runway implements new safeguards like an improved visual moderation system and C2PA provenance standards.
Gen-3 Alpha utilizes highly descriptive, temporally dense captions during training to enable imaginative transitions and precise key-framing of elements within a scene. The model excels particularly at generating photorealistic human characters with expressive faces and a wide range of actions, allowing creators to execute complex cinematic prompts with remarkable accuracy.