Runway’s new video-generating AI, Gen-3, offers improved controls
When will Gen-3 Alpha be available for Runway subscribers, and who are the intended users?
Gen-3 Alpha will be available for Runway subscribers in the coming days. The intended users are Runway subscribers, including enterprise customers and companies in Runway's creative partners program1.
What are the key improvements of Runway's Gen-3 Alpha model compared to its predecessor, Gen-2?
The key improvements of Runway's Gen-3 Alpha model compared to its predecessor, Gen-2, include:
-
Improved generation speed and fidelity: Gen-3 Alpha offers a "major" improvement in generation speed and fidelity over Runway's previous flagship video model, Gen-2. This means that it can generate videos with higher quality and in less time than its predecessor.
-
Fine-grained control over video structure, style, and motion: Gen-3 Alpha provides users with more precise control over various aspects of the videos it creates, such as the structure, style, and motion. This allows for more creative freedom and customization when generating videos.
-
Better interpretation of styles and cinematic terminology: Gen-3 Alpha has been designed to interpret a wide range of styles and cinematic terminology, enabling it to create more diverse and visually appealing videos.
-
Improved handling of human characters: Gen-3 Alpha excels at generating expressive human characters with a wide range of actions, gestures, and emotions, making it better suited for creating videos featuring human characters than its predecessor.
-
Shorter generation times: Gen-3 Alpha has reduced generation times compared to Gen-2. A 5-second clip takes 45 seconds to generate, and a 10-second clip takes 90 seconds to generate.
However, it is important to note that Gen-3 Alpha still has some limitations, such as a maximum footage length of 10 seconds and challenges with complex character and object interactions.
What specific capabilities does Gen-3 Alpha have in terms of generating human characters and cinematic effects?
Gen-3 Alpha is designed to generate expressive human characters with a wide range of actions, gestures, and emotions. It can interpret a wide range of styles and cinematic terminology, enabling imaginative transitions and precise key-framing of elements in the scene. The model can also generate video clips from text descriptions and still images.