Today, Google DeepMind introduced Gemini Omni 1.1 Flash, a new model with advanced creative controls and generative video capabilities for developers. The update makes the model production-ready via the Gemini API in Google AI Studio. Key features include scene extension, which allows the model to analyze up to 10 seconds of prior context (up from just the final second) to generate seamless continuations, enabling longer stories up to 40 seconds in total.

Developers can also specify first and last frames to achieve smooth transitions and camera movements, such as whip-pans or zooms, ideal for complex shots. For efficiency, Omni 1.1 Flash supports 360p preview generation at up to 60% faster speeds and a third of the cost compared to 720p, facilitating rapid prototyping. Additionally, upscaling to 1080p or 4K is available for professional production.

These tools empower developers to build generative video workflows, creative tools, and media editing applications with greater control and polish.