Google DeepMind made Gemini Omni 1.2 Flash generally available on 5 November 2026, across the Gemini API and Gemini for business. It is the first time the Omni video line has left preview and become a production product, backed by Google Cloud infrastructure and security — and the headline changes are the two things creators ask about first: clips are longer, and generation is cheaper.
Clips now run natively to 15 seconds, up from 10 seconds with Omni 1.1, with duration selectable in one-second increments from 3 to 15. The price for 720p video falls 30% to $0.07 per second, which puts a full 15-second clip at about $1.05. The workflow Google recommends keeps costs down further: draft fast at 360p or 720p, fix a detail with a follow-up edit instead of re-generating the whole clip, then upscale your best creations to 1080p or 4K.
Editing in conversation, not starting over
The most practical upgrade for anyone making video is multi-turn conversational editing. You refine the same clip turn by turn — swapping a product or background, changing the lighting or weather, switching the aspect ratio — instead of starting from scratch. Google says this works across full 15-second clips, so one generated scene can become ad variants for different audiences and placements without re-briefing or re-generating.
Google's own example: starting from one generated clip, follow-up prompts swap the background turn by turn — from a striped pavilion tent to an oak forest, a stable with a white horse, a market fair, and a stone castle — while keeping both characters' performances, armor reflections, and camera framing intact. That is Google's demo, not an independent test, but it shows what the feature is built for: iteration on a clip you already like, rather than rolling the dice again.
References that actually stick
Omni 1.2 Flash is a reference-driven model: you guide generation by combining character, product, or style images, short motion clips of up to three seconds, and audio cues — and Google says the model holds onto those details from start to finish. Faces, garments, packaging typography, and brand colors are meant to stay consistent across camera angles, which is what makes the model useful for commercial work rather than just one-off clips.
The showcase example is virtual try-on. Combining one model photo with front-and-back product shots of three outfits, Google says Omni 1.2 Flash generates a 15-second virtual try-on sequence across 360-degree turns and match cuts while keeping the model's identity and each garment's cut, drape, and back details consistent. The same mechanism turns existing product photography or character sheets into catalog spins and branded video ads — the kind of work a creator or small seller would previously need a shoot, a model, and an editor to produce.
Consistency is the other half of the claim. Google says character, object, and scene consistency hold from the first frame to the last, with less visual drift or morphing across motion and camera transitions — whether you are interpolating between a first and last keyframe or extending a clip to build a longer sequence. In Google's comparison, Omni 1.1 Flash fails to match a reference exhibition poster as closely as 1.2, which builds the poster layer by layer and locks every material, shadow, and typographic detail through the final frame.
More natural speech and a steadier camera
Two more upgrades target the things that make AI video look AI-made. Dialogue is meant to sound more fluid and natural, with clearer lip-sync and distinct multi-speaker conditioning — each sampled voice routed to the right character — whether you are directing multi-character scenes in Flow, making explainers in Gemini for business, or producing avatar videos in Google Vids. And camera work is more controllable: complex, continuous movements like sweeping FPV flythroughs, dramatic zoom-outs, and orbits come with significantly fewer unwanted jump cuts. Google's demo is a single continuous FPV drone flight across a piazza, through bronze doors, and up through the Pantheon's coffered dome to the oculus — with, Google says, zero jump cuts.
Where you can use it
Omni 1.2 Flash (API model ID gemini-omni-1.2-flash) is available now across Google's developer and consumer surfaces:
- Developers: the Gemini API, prototyped in Google AI Studio, with Provisioned Throughput for high-volume production workloads.
- Businesses: the Google Cloud console, Gemini for business, and Google Ads for video campaigns at scale.
- Everyone: Flow for directed scenes and multi-turn edits, Google Vids — where 15-second generation is available at no additional cost with a Google account or Workspace subscription — and the Gemini app, where 15-second generation is rolling out to Google AI Ultra subscribers.
Every output carries an imperceptible SynthID watermark, and the model is production-ready with enterprise SLAs. Matt Rouif, CEO of Photoroom, is quoted in Google's Cloud announcement saying: "Version 1.2 brings 15-second videos, a top user request. With general availability and SLAs, we can rely on it in production for our sellers and enterprise customers."
The limits Google admits
Google is candid about what is still weak: fast sports action, rapid script changes, and complex physics remain difficult, and the company says development there is ongoing. That honesty matters, because the consistency and quality claims above come from Google's own demos and write-up — no independent lab has published hands-on results yet. The 30% price cut and the 15-second cap are verifiable facts; "stronger consistency" and "more natural speech" are claims to test against your own footage.
Why it matters for creators
Fifteen seconds is a meaningful threshold. Ten-second clips are teasers; fifteen seconds is enough for a complete story beat, a product explainer, or a full social ad spot in a single generation — no stitching, no mid-sentence cut. Combined with a price that makes a 720p draft cost about a dollar, and editing that lets you fix one detail instead of re-rolling the whole clip, Omni 1.2 Flash is the first version of Google's fast video model that reads as a production tool rather than a preview.
For the creators reading this blog, the test is straightforward: take a product photo or a character sheet you already have, generate a 15-second clip, and see whether the references survive to the last frame. Google says they do. The model is generally available now, so you can check that claim yourself today.
Sources and classification
Classification: Official Announcement (Google DeepMind, 5 November 2026). Launch facts, pricing, and capability claims are from Google's announcement and Cloud blog; quality and consistency claims are Google's own, with no independent verification published yet. Checked 7 November 2026.
- Google Blog, "Gemini Omni 1.2 Flash: 15-second video and lower pricing" (official announcement by Anish Nangia, Product Manager, Google DeepMind, 5 November 2026): https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/
- Google Cloud Blog, "Gemini Omni 1.2 Flash now GA: cost-effective video" (6 November 2026) — GA confirmation, Provisioned Throughput, Photoroom quote: https://cloud.google.com/blog/products/ai-machine-learning/gemini-omni-1-2-flash-generally-available
- Google Cloud documentation, Gemini Omni 1.2 Flash model page (API model ID gemini-omni-1.2-flash; generation, editing, extension, upscaling): https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/gemini/omni-1-2-flash
- Tech Times, "Google Debuts Gemini Omni 1.2 Flash With Longer AI Clips, Better Editing, and Fewer Errors" (6 November 2026) — launch context and availability: https://www.techtimes.com/articles/330142/20261106/google-debuts-gemini-omni-12-flash-longer-ai-clips-better-editing-and-fewer-errors.htm