Gen-3 Alpha is Runway's latest text-to-video AI model, representing a significant advancement in video generation technology.
High-fidelity video generation: It produces photorealistic human characters and intricate environments with nuanced actions and emotions.
Improved consistency and motion: Gen-3 Alpha demonstrates superior temporal consistency, maintaining stable and coherent characters and elements throughout the video.
Fast generation: It can produce 5-second videos in about 60 seconds and 10-second videos in about 90 seconds.
Fine-grained temporal control: The model allows for precise control over video generation, enabling smooth transitions and exact placement of elements in the video timeline.
Multimodal training: Gen-3 Alpha is trained on a new infrastructure for large-scale multimodal learning, combining video and image training.
Versatile applications: It powers various Runway tools, including Text to Video and Image to Video.
Advanced prompting: Gen-3 Alpha responds well to detailed prompts, including descriptions of subjects, scenes, lighting, camera movements, and actions.
Resolution and frame rate: Currently, it generates videos at 720p resolution and 24 frames per second.
Gen-3 Alpha is now available to users with a paid Runway account, with subscription plans starting at $15 per month. It represents a step towards building General World Models capable of understanding and generating complex visual content.