A ten-second clip is long enough for a product shot, but often too short for the shot plus a reaction or a second camera move. Google’s new video model targets exactly that gap. On 5 November 2026 Google released Gemini Omni 1.2 Flash, which creates and edits native video of up to 15 seconds and costs 30 percent less than its predecessor.
What changed from Omni 1.1 to 1.2 Flash?
Omni 1.1 Flash was limited to 10 seconds. Version 1.2 goes up to 15 seconds, and you can set the length in one-second steps from 3 to 15 seconds. Google calls it the first generally available version of Gemini Omni. According to Google, the model also follows reference material more closely and keeps characters, objects and scenes more consistent from the first to the last frame, with less drift or morphing. Camera work is supposed to be calmer too: fewer unwanted jump cuts, steadier scale and longer continuous moves, such as an uninterrupted flight shot. For dialogue, Google promises clearer lip-sync and multi-speaker conditioning, meaning several voices can be assigned to the right characters within one scene.
What does a clip cost now?
On the Gemini Developer API, Omni 1.2 Flash (model ID gemini-omni-1.2-flash) costs about $0.07 per second at 720p, so roughly $1.04 for a full 15-second clip. Omni 1.1 Flash was listed at about $0.10 per second. The pricing page also lists about $0.023 per second at 360p, and about $0.104 (1080p) or $0.209 (4K) per second with upsampling. There is no free API tier for the model. Google recommends drafting at 360p or 720p, refining with follow-up edits and only then upscaling.
How does conversational editing work?
Instead of writing a new prompt and regenerating everything, you keep talking to the model about the same clip. Google describes multi-turn editing across the full 15 seconds: swapping a product or background, changing lighting or weather, or switching the aspect ratio. As references, the model accepts character, product or style images, short motion clips of up to three seconds and audio cues. Google says faces, clothing, packaging typography and brand colours should stay consistent across camera angles.
Where can you use it?
The model is available on the paid tier of the Gemini API, for prototyping in Google AI Studio and for businesses through Google Cloud and Gemini for business. In Flow you can direct scenes and edit them over several turns. Google Vids generates 15-second clips at no additional cost with a Google account or Workspace subscription. In the Gemini app, 15-second videos are rolling out to Google AI Ultra subscribers, and Google Ads uses the model for video campaigns.
What happens to Omni 1.1 Flash users?
From 3 December 2026, requests to gemini-omni-1.1-flash will be routed automatically to the 1.2 model and billed at its rates. If you build on the API, it is worth re-checking your existing prompts and clip lengths before that date.
Limits and safeguards
Google itself names fast sports action, rapid script changes and complex physics as areas still under development. Every output carries an invisible SynthID watermark that marks the video as AI-generated.
Our assessment
The documented facts are the longer clip, the lower per-second price and the automatic switch in December. The quality claims about consistency, lip-sync and camera motion come from Google’s own examples, not from independent tests. A striking frame. What about the motion? That is the question to answer with your own scene: generate the same shot with the same reference several times and check separately whether the character holds and whether the camera follows your instruction. For teams producing many short variants, the combination of lower draft prices and editing without regeneration is likely the more practical gain than the extra five seconds.
