Google cuts Gemini Omni video prices about 30% and makes the model generally available

Google on November 5, 2026, released Gemini Omni 1.2 Flash, a new version of its model for generating and editing video, and says the Omni line is generally available for the first time. Developers on the paid tier of the Gemini API now pay about $0.07 per second of 720p video, against about $0.10 for the previous version, and one generation can run 15 seconds instead of 10.

The update comes less than six months after the model's debut. Google put the first Omni model into the Gemini app, Google Flow and YouTube Shorts in May, opened it to developers on June 30 and, with version 1.1 on August 27, added scene extension, control over a shot's first and last frames, and 4K output.

What the cut covers

Only the video output rate moved. Google's price list shows it falling to $12.00 per million tokens from $17.50, while input stays at $1.50 per million and text output at $9.00. That is a 31% drop, by our arithmetic; Google describes it as 30%. Neither version has a free tier on the API.

On those rates, Google puts a 15-second 720p clip at about $1.04. A 10-second 720p clip, the longest the old version could make in one generation, drops to about $0.70 from about $1.01; the older figure is our calculation from the listed token rate.

Developers who stay on the old model will not keep it for long. The price list says that from December 3, 2026, requests to Omni 1.1 Flash will be routed to 1.2 and billed at its rates. As we read that notice, applications still calling the older version will get the newer model's output from that date, whether or not their owners have retested them.

What Google says is new

Clip length can now be set in one-second steps from 3 to 15 seconds, and follow-up instructions can change a background, the lighting or a product in an existing clip without starting again. Google also says the model holds faces, clothing, packaging text and brand colors steadier across camera moves when given reference images, and that speech and lip-sync sound more natural. These are the company's descriptions. Its post shows example clips but gives no benchmark scores.

Google's technical documentation sets limits the announcement does not dwell on. Clips built between a specified first and last frame still top out at 10 seconds, and extending a video adds 3 to 10 seconds at a time. The Google Cloud post carries endorsements from companies such as Photoroom and Picsart; those are customers' statements, not independent tests.

development is ongoing to further improve things like fast sports action, rapid script changes and complex physics

Watermarks and availability

Google says output carries its invisible SynthID watermark, one of the methods in our explainer on how AI content labels and watermarks work. Its model page adds C2PA Content Credentials for material produced or edited with Omni in the Gemini app, Flow or YouTube.

Beyond the API and Google Cloud, the company says the model is rolling out in Flow, in Google Ads and in Google Vids, where it says 15-second clips come at no additional cost with a Google account or Workspace subscription. Google AI Ultra subscribers are getting it in the Gemini app.

What comes next

The December 3 switch is the next fixed date. Google's Cloud documentation gives November 5, 2027, or later as the date support for version 1.2 ends on that platform. In May, Google said the Omni family would in time add image and audio as outputs; the new post gives no date for that.