Google announced the release of the Gemini Omni 1.1 Flash video generation AI model, capable of outputting videos with a maximum resolution of 4K. Compared to the standard 720p of Omni 1.1, the new model makes notable trade-offs in speed and cost, extending towards professional production scenarios.
The full workflow from storyboard preview to 4K final video
One of the highlights is scene expansion: users can seamlessly continue generating subsequent frames based on an existing video, expanding step by step in 10-second increments, with a maximum of 40 seconds per clip. The specified start and end frame feature allows smooth transitions and camera movements by simply providing the starting and ending frames; video reference support allows introducing up to a 3-second segment, maintaining precise context and character consistency.
The cost structure is clearly layered: 360p preview costs $0.03 per second, with speed up to 60% faster than standard and only one-third of the cost, suitable for storyboard iteration; 720p, 1080p, and 4K cost $0.1, $0.15, and $0.3 per second respectively. Google is moving video generation from "previewable" to "deliverable," creating a continuous production pipeline between prototypes and final versions for AI videos.
