Runway, an AI video startup, has announced its latest model, Gen-4, which aims to solve several key problems with AI video generation. The main advantage of Gen-4 is its ability to maintain consistent characters and objects across shots, a feature that was previously lacking in previous models.
In Gen-4, users can provide a single reference image of the character or object they want to see consistently, and the model will generate videos with accurate representations throughout. For example, Runway has released example videos showing the same woman appearing in various different shots across different scenes, while maintaining her appearance.
Another significant improvement is Gen-4’s ability to capture multiple angles within a single scene, allowing filmmakers to get coverage of the same environment or subject from different perspectives. This feature was previously only available in limited form with previous models.
Gen-4 builds upon Runway’s existing model, Gen-3, which was announced last June and expanded video length from 2 seconds to 10 seconds while offering greater consistency and coherence. With the release of Gen-4, Runway has taken another step towards improving its AI-powered video synthesis tool, positioning itself as a major player in the crowded space of AI-generated content creation.
Source: https://arstechnica.com/ai/2025/03/with-new-gen-4-model-runway-claims-to-have-finally-achieved-consistency-in-ai-videos