Google announces Gemini Omni 1.2 Flash with the ability to make 50% longer video clips

Affiliate links on Android Authority may earn us a commission. Learn more.

Google announces Gemini Omni 1.2 Flash with the ability to make 50% longer video clips

Nov 5, 2026 — 3:11 PM ET

- Google has announced Gemini Omni 1.2 Flash, its latest AI video generation model, with support for clips up to 15 seconds long.

- The new model promises better visual consistency, improved reference adherence, more natural speech, and multi-turn conversational editing.

- Google AI Ultra subscribers get access through the Gemini app, while free users can try some of the new capabilities through Flow and Google Vids.

Google’s family of video generation models is getting a new addition, and this one brings some solid improvements at a cost-effective price point. The new model, Gemini Omni 1.2 Flash, introduces native 15-second video creation, multi-turn conversational editing capabilities, better reference following (compared to older models), and overall more consistent results.

Although truly getting started with Omni 1.2 Flash through the Gemini app will initially require a Google AI Ultra subscription, there’s a little something for free users, too.

One of the biggest improvements is the ability to generate longer videos. Gemini Omni 1.1 Flash, for reference, was limited to 10-second generations. Omni 1.2 Flash increases that limit to 15 seconds, giving you 50% more time to work with. The model also offers control over video length — you don’t necessarily have to generate 15-second videos — you can specify durations between three and 15 seconds in one-second increments.

According to Google, the new model offers better reference adherence too. So if you’re planning on generating multiple clips to tell a story, the new model should do a better job than its predecessors when it comes to keeping characters and other elements consistent across scenes. “You can maintain stronger character, object, and scene consistency from the first to the last frame, with less visual drift or morphing across motion and camera transitions. Whether you’re interpolating between a first and last keyframe, or extending a clip to build a longer sequence, products, characters, props, and environments stay consistent from shot to shot,” wrote the tech giant.

Another major upgrade that Google highlighted was easier result editing. Gemini Omni 1.1 Flash, for reference, supports conversational video editing. Omni 1.2 Flash takes that a step further by allowing turn-by-turn changes without having to start over from scratch, as seen in the image above. Google says this should make it much easier to produce different variants of a video for different audiences, all without the need to re-brief the tool or re-generate the video from scratch.

During all changes, Omni 1.2 Flash promises to preserve the elements you don’t want changed.

Lastly, expect better speech generation with more fluid and natural-sounding dialogue with proper lip-syncing, alongside more control for cinematic projects, like sweeping FPV flythroughs and dramatic zoom-outs.

Availability

Gemini Omni 1.2 Flash is making its way to several Google apps, though access doesn’t look the same across the board.

Users will be able to use the Gemini app to create 15-second videos, but only if they’re a Google AI Ultra subscriber. Developers can access the model via the Gemini API, while businesses can take Omni 1.2 Flash for a spin in the Google Cloud console.

Lastly, for free users, the new model is also coming to Flow. Google hasn’t explicitly said whether free users will have access to all Omni 1.2 Flash capabilities in Flow, but it did say that they’ll be able to “Direct scenes and multi-turn edits.” Similarly, Google Vids gets access to 15-second video generation at no additional cost, though it’s unclear if other Omni 1.2 Flash improvements are making their way to free Vids users.

It’s plausible that the new model will eventually be available to Google AI Plus and Pro users. When, though, remains uncertain.