Google's video model can now be told where a shot starts and where it ends — and drafts cost a third as much
Google released Gemini Omni 1.1 Flash on Thursday 27 August 2026, a production update to its video generation model aimed at people building tools and workflows rather than at one-off clips. Four things are new. Scene extension lengthens a clip in 10-second increments up to 40 seconds, and now reads up to 10 seconds of preceding footage for context instead of only the last moment, which is what keeps the extension visually consistent. Keyframe control lets you supply a first and last frame and have the model generate the transition and camera move between them. A 360p preview mode produces rough drafts up to 60 percent faster and at a third of the cost of standard 720p output, so an iteration loop stops being expensive. Video references let a prompt cite up to three seconds of existing footage to hold a character or a look steady across shots. Finished work can be upscaled to 1080p or 4K. It is available in Google AI Studio, through the Gemini Enterprise Agent Platform API, in Google Flow for Plus, Pro and Ultra subscribers, and scene extension is coming to the Gemini app on paid tiers.
Related stories
Google's new transcription model does not write down what you said — it writes down what you meant, in 85 languages2026-08-28Ten million hours of video, released for anyone to train on — the largest open video set yet comes from a German non-profit
2026-08-27The stealth model developers spent weeks probing is now MIT-licensed and downloadable — and Z.ai says the entire preview ran on Chinese chips
2026-08-27IBM's new open models were not taught to describe using a terminal. They were trained by using one.
2026-08-27The nameless model that burned through 42 trillion tokens in six days has an owner: Z.ai says Ox Alpha is a GLM, and the weights are opening
2026-08-26