Google announced Gemini Omni 1.2 Flash on November 5, 2026, raising the length of videos it can generate and edit from 10 to 15 seconds. For creators and small businesses making product explainers or social ads, that means more room for a message in a single generated clip.
The model is now generally available through the paid Gemini API—the service used to build it into other software. But the lower API price does not mean every Google video app costs less, and access to longer clips depends on the product and plan.
What a 15-second clip costs
Google’s pricing table lists video output at $12 per million tokens, down from $17.50 for Omni 1.1. At 720p, Google counts 5,792 video tokens per second; these are billing units.
For one 15-second result, the calculation is 15 × 5,792 × $12 ÷ 1,000,000 = about US$1.04. Three completed attempts would cost about $3.13 in video output alone—an illustrative budget, not a measured success rate. Input processing, other billable output, further edits and human review are outside those figures.
In Flow, billing is separate: a 15-second 720p generation uses 60 credits (10 × 3, plus 5 × 6), compared with 30 credits for 10 seconds on Omni 1.2. That is 50% more duration for twice the credits. Credits are not dollars; the API figure does not price a Flow subscription.
Which apps offer the longer videos?
Flow: Clips of up to 15 seconds are listed for Google AI Plus, Pro and Ultra subscribers; generation and editing up to 10 seconds are listed for all users.
Gemini app: Google says 15-second creation is rolling out to Google AI Ultra subscribers.
Google Vids: Google announced 15-second clips at no additional cost with a Google account or Workspace subscription. Its help page still describes eligible subscriptions and 3–10-second controls, so universal access is not yet established.
The extra seconds have limits
The API supports video with synchronized sound and follow-up editing instructions. However, setting both the opening and closing image still limits a clip to 10 seconds. A creator who needs to define both endpoints therefore faces a different choice from someone who needs a full 15-second scene.
For uploaded footage, the Gemini API’s regional restrictions also matter: editing and extension are unavailable in the European Economic Area, Switzerland and UK. Editing or extending model-generated clips remains supported there.
Google claims better visual consistency and speech synchronization, but acknowledges work remains on fast action and complex physics. For a product promotion, check packaging text, the product’s appearance and movement before using the result; those quality claims do not guarantee usable footage.
Google says the videos contain an invisible SynthID watermark. As our guide to checking AI media explains, that helps identify AI involvement; it does not establish that an event or product claim is true.