Google's Gemini Omni 1.2 Flash Doubles Down: 15-Second AI Video at 30% Less

Google cut pricing on Gemini Omni 1.2 Flash by 30% and extended native video generation from 10 to 15 seconds, the company announced on November 5, 2026. The model is generally available through the Gemini API and Gemini for business, with integrations spreading across Flow, Google Vids, the Gemini app and Google Ads.

That's the headline. Now the details — and for advertisers and e-commerce teams, they matter.

What does the new version actually change?

Users can specify video durations in one-second increments from three to 15 seconds. Google positions the flexibility for advertisements, product demonstrations, educational content and short-form video.

The bigger cost lever is workflow. Users generate initial drafts at 360p or 720p, then upscale selected outputs to 1080p or 4K. Instead of regenerating an entire video when something changes, users refine existing clips through conversational editing — adjusting backgrounds, products, lighting, weather conditions and aspect ratios while keeping the underlying sequence intact. Google says this cuts the computing costs of repeated generations.

One sentence: draft cheap, polish only what survives.

How strong is brand consistency?

Reference adherence is the upgrade most likely to pull in commercial buyers. The model accepts character images, product photography, stylistic references, short motion clips up to three seconds long, and audio cues as guidance.

Google says the model maintains consistency in facial features, clothing, packaging, typography and brand colors across generated videos — the details that usually break in AI video and force manual fixes. The company highlighted a demonstration: a 15-second virtual try-on sequence built from a model photograph and product images of several outfits.

The model also holds consistency across camera angles and transitions, reducing unwanted changes to characters, objects and environments.

What about speech and camera control?

Speech generation got three upgrades: more natural dialogue, improved synchronization between speech and facial movements, and more distinct voices for multiple speakers. Google targets AI-generated explainers, marketing videos, character dialogue and virtual presenters.

Camera control expanded too. Users can generate complex continuous sequences — first-person flythroughs, dramatic zoom-outs, orbiting camera movements — with fewer unwanted cuts and more consistent visual scale. Google demonstrated the technology with architectural environments, animated sequences, and motion transfer between subjects.

Who is already using it?

Google named four organizations working with the model in production:

• Adobe plans to incorporate the updated model into Firefly, giving users longer video outputs and improved consistency with reference images.
• Monks, a creative production company, has been testing the model for visual effects and controllable video generation.
• Havaianas is among the early adopters Google identified.
• Whering, a fashion technology company, has evaluated how accurately the model represents clothing and fabrics in virtual styling applications.

Enterprise adoption runs on Google Cloud infrastructure, including Provisioned Throughput for organizations that need predictable computing capacity. Every generated output carries a SynthID watermark — imperceptible to viewers, but present — to identify AI-generated content.

How do you get access?

Developers can use Google AI Studio and the Gemini API. Businesses go through Google Cloud. Google Vids users can generate 15-second clips at no additional cost with a Google account or Google Workspace subscription. Access through the Gemini app is rolling out to Google AI Ultra subscribers.

What are the limits?

Google acknowledged the model still struggles with fast-moving sports footage, rapid changes in scripted scenes, and complex physical interactions. Those gaps keep it out of some premium production contexts for now.

With longer clips, 30% lower pricing and editing that doesn't require full regeneration, Gemini Omni 1.2 Flash is Google's clearest push yet to move generative video from demo territory into paid advertising, entertainment and e-commerce workflows at scale.