Google's new video model prices a 360p draft at about a third of standard cost, letting marketing teams iterate on ad concepts before paying for a final render.
Every marketing team has lived through some version of this loop: brief a production vendor on a fifteen second product spot, wait three or four business days for a first cut, sit through a review call, send back two pages of notes about pacing and that one transition at the eight second mark, then wait again for the next round. On August 27, Google shipped an update aimed at collapsing that loop into something you can run yourself, in minutes, before a vendor ever sees the brief.
Gemini Omni 1.1 Flash is Google's latest production ready release of its video generation model, and the feature that matters for anyone running creative operations is a 360p draft mode. It renders up to 60 percent faster than the model's standard 720p output and, according to Google's pricing page, costs roughly a third as much. Standard 720p output runs about ten cents a second under the API's list pricing. A 360p draft of the same clip lands closer to three cents a second, cheap enough to generate four or five competing concepts before you decide which one is worth finishing.
The other real change is control. Omni 1.1 can now hold onto ten seconds of a video's prior context when extending a scene, up from roughly one second in the previous version, and you can chain extensions in ten second increments up to a cumulative 40 seconds. You can also specify an exact first frame and last frame and let the model generate everything in between, which is the difference between describing a camera move to an editor on a call and just building the move directly. Finished cuts upscale to 1080p or full 4K for delivery.
None of this is theoretical launch talk. Adobe has built Omni Flash into Firefly. Figma's Weave canvas uses it as one of its video models, with Figma creative director Itay Schiff telling Google that the added resolution and reference controls take teams "beyond generating videos to truly directing them." Runway's chief creative officer described it as a tool that slots into how creators already work: start with a prompt, an image, or a video, then generate or edit from there. Omni 1.1 is live now in Google AI Studio and the Gemini Enterprise Agent Platform, and the update has also reached Google AI Plus, Pro, and Ultra subscribers inside the Gemini app and Flow.
Here's the part that actually changes a marketing operation. The expensive piece of ad video has never really been the render, it's been the number of rounds you can afford to iterate before committing budget to a real shoot or a finished cut. When a draft costs three cents a second instead of ten, and finishes in a fraction of the time, you can afford to be wrong four times before you're right once. That turns creative review from a gate you pass through once, expensively, into something closer to a group chat. A CMO does not need to understand tokens per second to understand what happens to a production timeline once the exploratory phase gets an order of magnitude cheaper.
The honest caveat: this is a drafting and iteration tool, not a replacement for a shoot with your actual product, your actual spokesperson, or a real location. Google's own developer documentation lists real limits worth knowing before you plan a campaign around this. You cannot upload footage of someone talking and add new dialogue to it, only silent footage can be extended with speech layered in later through multi-turn generation. Editing or extending uploaded video is not available at all for users in the European Economic Area, Switzerland, or the UK. Extensions can only be appended to the end of a clip, never inserted in the middle. And every output carries an invisible SynthID watermark, the right call for provenance, but a reminder that this is generated media, not a substitute for footage of your product on an actual shelf.
So no, this does not put the production shop out of business the way some of the launch framing wants you to believe. What it does is move the expensive commitment point later. You used to have to guess right on the first pass, because guessing wrong cost real money and real days. Now you can guess wrong nine times for what one guess used to cost, and only bring in the crew once you already know which sixty seconds are worth shooting for real.