Google's Gemini Omni 1.2 Flash makes 15-second video clips and cuts the price 30% from Omni 1.1

Google released Gemini Omni 1.2 Flash, the latest version of its video generation and editing model, on 5 November 2026. Its announcement post says the model makes clips of up to 15 seconds natively and costs 30% less than Gemini Omni 1.1.

In August we covered the previous version, under the headline Google's new Gemini video model is built for developers, not the feed. Google's Cloud post pitches 1.2 at businesses making video in volume, naming "product videos, virtual try-ons, and ads".

What Google says is new against 1.1

• Longer clips. Up to 15 seconds, "up from 10 seconds with Omni 1.1". You can set the length in one-second steps from 3 to 15 seconds.
• Closer reference following. Google says faces, garments, packaging text and brand colours stay consistent across camera angles. One of its demo captions says that "Gemini Omni 1.1 Flash does not match the reference image as closely as 1.2 does".
• Editing in conversation. You can change the same clip turn by turn, such as swapping a background or the lighting, across full 15-second clips.
• Speech and camera work. Google describes more natural dialogue, clearer lip-sync, and continuous camera moves with "significantly fewer unwanted jump cuts".

One inconsistency: the main post says you can guide a clip with "audio cues". The Cloud post says "Audio references are gated behind allowlisting", and the Gemini API documentation says "Uploading audio inputs or audio references is unsupported in the current version of the API."

The price, and what the 30% is measured against

Google's comparator is its own previous model: "We have dropped the price 30% from Gemini Omni 1.1 making it available at $0.07 per second for 720p video." The Cloud post adds "with similar savings at other resolutions".

The Gemini API pricing page bills video by output tokens, at $12.00 per million for 1.2 against $17.50 for 1.1. On our arithmetic that is a cut of about 31%. Google's own per-second figures for 1.2 on the paid tier (there is no free tier):

• 360p: $0.023 per second
• 720p: about $0.07 per second (about $0.70 for a 10-second clip, $1.04 for 15 seconds)
• 1080p with upsampling: $0.104 per second
• 4K with upsampling: $0.209 per second

For 1.1 the page gives only the 720p figure, about $0.10 per second. From 3 December 2026, all API requests to 1.1 will be routed to 1.2 and billed at 1.2's rates, the page says.

Who can use it

• Developers: the model is generally available on the paid tier of the Gemini API and on Google Cloud. Google Cloud's model page lists its region as "global".
• Gemini for business: the main post says the model is generally available there. The Cloud post says it is "coming soon to Gemini for business".
• Google Vids: anyone with a Google account or Workspace subscription can make "up to three 15-second videos per month in Vids at no cost", Google's Workspace post says.
• Gemini app: 15-second videos are "rolling out to Google AI Ultra subscribers". Google also lists Flow and Google Ads among the products it is rolling out to.

Google's API documentation names the UK among the places with limits. Editing or extending uploaded videos is "not currently available" in the European Economic Area, Switzerland and the UK. Uploading and editing images that contain minors is also not supported in those places. None of the posts says when the consumer products will reach UK users.

Safety and limits, as Google states them

• "Outputs carry an imperceptible SynthID watermark", the Cloud post says. The documentation adds that the watermark "can be detected programmatically".
• Google Cloud's model page lists Content Credentials (C2PA) as supported.
• Content safety filters check both prompts and output, and those filters "vary by region". Uploading images of "certain recognizable people" is not supported.
• Google says work is still under way on "fast sports action, rapid script changes and complex physics".

Benchmarks

None of the three Google posts we read in full compares 1.2 with any rival model. There are no benchmark tables and no charts. The only image in the main post is its title graphic. Its comparisons are demo videos set against 1.1, and we have not reviewed them frame by frame. Of the customers quoted in the Cloud post, invideo said that after its own evaluations the model is "one of the best we've tested" for fast video at scale.

Why it matters

On our reading, the bigger change is the bill rather than the clip length: Google says the 30% saving at 720p is matched by "similar savings at other resolutions". Once Gemini API requests to 1.1 are routed to 1.2 on 3 December, those requests are billed at 1.2's rates. Also on our reading, the UK limits on editing uploaded footage bear on the product-video work Google is pitching the model for, so UK developers would do well to read the documentation first.