Google made Gemini Omni 1.2 Flash generally available on Google Cloud today. It’s the newest version of Google’s video generation and editing model, and Google is positioning it for the videos businesses make over and over: product videos, virtual try-ons, and ads. Developers can use it now through the API and the Google Cloud console. Business users can create with it in Google Vids now, and in Gemini for business soon.
This release is less about wow-factor demos and more about the economics of video production. Prices are lower, clips are longer, and you can change a clip without starting over.
What’s new in 1.2
- Lower price. 720p video costs $0.07 per second. Google says that’s 30% less than Omni 1.1 Flash, with similar cuts at other resolutions.
- Longer clips. Generate and edit videos up to 15 seconds long. The old limit was 10.
- Closer reference following. The model sticks more closely to image and video references, so products, characters, and brand details look the same from shot to shot.
- Draft cheap, then upscale. Make drafts at 360p or 720p, then upscale the keepers to 1080p or 4K.
- More natural speech. Dialogue flows better, lip-sync is clearer, and multiple speakers can be conditioned separately.
- Editing by conversation. Refine the same clip by swapping a product, changing the lighting, or moving the camera angle, instead of generating it again from scratch.
One catch: audio references are allowlist-only for now.
Why it matters for agencies and in-house creative teams
The 15-second limit is the change people will notice. A full scene fits in one generation, and that covers a typical short social ad. Before, you’d have to stitch two clips together.
The cost math is simple. At the announced rate, a 15-second clip at 720p comes to about $1.05 before upscaling. That makes it practical to generate many versions of a concept, keep the best ones, and spend upscaling money only on finalists.
Reference adherence matters most for brand work. If the product, the logo, or the spokesperson changes from shot to shot, the clip is useless to a client, no matter how good it looks. Google is betting that better consistency plus conversational edits will move these models from experiments into real production pipelines.
Who’s already using it
Google named several creative platforms that are building Omni 1.2 Flash into their products:
“Gemini Omni 1.2 Flash brings exciting capabilities in reference-based generation, video extension, and creative consistency, helping our customers move from experimentation to professional content production at scale.”
Roee Peled, CPTO, Artlist
“Gemini Omni 1.2 Flash edits exactly what you ask for, a colorway or a background, and leaves the product untouched, even from detailed images and finished video.”
Andrew (Taco) Zhang, Product Marketing Manager, FLORA
“It sticks to the prompt, and the physics and sound feel real. If you need fast video at scale, it’s one of the best we’ve tested. And 15 seconds means a full scene in one go.”
Shivam R., VP of Engineering, invideo
How to try it
- Developers: Test it in Vertex AI Studio in the Google Cloud console, build with the API, and check pricing.
- Business users: Try it in Google Vids now. Gemini for business is coming soon.
New to Google’s model family? Start with our explainer, What Is Google Gemini?
The Tygart take
Video generation is now competing on unit cost and controllability, not just visual quality. For teams that make video in volume, the workflow worth testing is: draft at 360p, refine through conversation, upscale only the winners. Run your own brand assets through it before you commit. Reference adherence is the claim that decides whether this is actually ready for production.
Source: Google Cloud announcement.
Get The Signal
What we actually ran, what it measured, and what it means for a restoration shop — every week, including the flops.
Leave a Reply