Gemini Omni 1.1 Flash extends video to 40 seconds

Google DeepMind announced Gemini Omni 1.1 Flash, positioning the video model as "production-ready for professional use" via the Gemini API in Google AI Studio.

The headline change is scene extension. The model now analyzes up to 10 seconds of prior footage before continuing a clip, compared with only the final second in previous models, which Google says improves visual consistency and narrative adherence. Extensions run in 10-second increments up to a cumulative 40 seconds.

Other additions are aimed at iteration cost and finishing quality. Developers can specify first and last keyframes and have the model generate continuous video between them, for camera orbits, zoom transitions, or looping clips. A 360p draft mode generates previews up to 60% faster — based on system throughput versus 720p — and at a third of the cost of standard 720p. Finished work can be upscaled to 1080p or 4K. Multimodal input now accepts up to three seconds of reference video for maintaining character and visual consistency.

Availability is immediate across Google AI Studio, the Gemini Enterprise Agent Platform, and Google Flow for Google AI Plus, Pro, and Ultra subscribers globally. Scene extension is also live in the Gemini app for those subscribers.

Google's post links a pricing table but does not state per-second or per-token rates in the text, and publishes no benchmarks, human-preference results, or comparisons against prior Omni versions or rival video models. The claims about consistency and speed are the company's own.