Google's Gemini Omni 1.2 Flash Extends Video to 15 Seconds, Cuts Price 30%
The Development
Google DeepMind made Gemini Omni 1.2 Flash generally available, rolling it out through the Gemini API, Google AI Studio, Google Cloud Console, Flow, Google Vids, the Gemini app for Google AI Ultra subscribers, and Google Ads. Clips now run three to 15 seconds, up from a 10-second cap, and pricing falls 30% to $0.07 per second at 720p. Conversational Editing lets users change backgrounds, lighting, or aspect ratio without regenerating the clip. The model accepts character images, product photos, motion clips of up to three seconds, and audio cues as inputs, with faces, clothing, packaging typography, and brand colors held consistent across camera angles. Every output carries an imperceptible SynthID watermark.
Our Take
The 30% price cut is the headline; the distribution is the story. Putting Gemini Omni 1.2 Flash inside Google Ads means video generation sits at the point of campaign creation, where the budget decision is made. The tiered workflow of 360p drafts, targeted edits, and a single 1080p or 4K upscale also changes cost structure. Cost per approved asset, not cost per generated second, becomes the metric that matters, and iteration is now nearly free. Reference adherence is the real unlock for brands. Drift in logos and packaging typography is what kept generative video out of paid media. Google claims that problem is largely solved. That claim needs testing against your own assets before you scale.
What Changed
Iterative, reference-locked video editing now exists at commodity prices. Marketers can feed product shots and brand assets into a model, hold identity and packaging fidelity across a 15-second sequence, and revise through conversation rather than re-rendering. Draft cheaply at low resolution, upscale only the approved cut.
Marketing Impact
Performance creative and ecommerce teams feel this first. Product-feed video, virtual try-on sequences, and per-audience variants shift from agency production cycles to in-platform generation. Creative operations must add reference-asset libraries and automated brand-compliance review, because volume will outrun manual approval.
Competitive Implication
Google Ads, Google Cloud, and Google Vids users gain a bundled workflow competitors cannot match without ad-platform distribution. Standalone video-generation vendors and production houses selling short-form product video are exposed: the model is cheaper, embedded where spend happens, and editable by conversation instead of by re-brief.
Strategic Outlook
Expect Meta Platforms and TikTok to answer with comparable native generation inside their ad tools within two quarters. Clip length and per-second pricing will keep compressing through 2027. The differentiator shifts from generation quality to reference fidelity, brand governance, and how fast teams test variants against live performance data.
The Exploit
Action Item
Performance creative leads should run a two-week Q4 test now: feed your top 20 SKUs' product shots into Gemini Omni 1.2 Flash via Google Ads, compare 15-second outputs against current video on cost per approved asset and ROAS, then set Q1 2027 production budgets accordingly.