Google's new video AI can extend a scene, jump to 4K, and cost less to test

Gemini Omni 1.1 Flash lands in the Gemini API with scene extension, keyframe control, cheap 360p drafts and 4K upscaling. What it costs and who actually pays.

AI2Day NewsdeskAI-assistedPublished Updated Editor: Lee Brown4 min read
A wide-angle editorial photograph of a modern open-plan office at dusk, rows of glowing monitors showing lines of code and numerical dashboards, warm overhead l
Illustration made with AI. Not a photograph of the events described.
Share

Key points

  • Google DeepMind launched Gemini Omni 1.1 Flash, a generative video model available through the Gemini API in Google AI Studio.
  • The model can extend an existing clip in 10-second chunks up to a total of 40 seconds, using the previous 10 seconds as context.
  • Draft videos at 360p resolution run up to 60% faster and cost about a third of the 720p price, according to Google.
  • Paying Google AI Plus, Pro and Ultra subscribers get scene extension inside the Gemini app and Google Flow starting today.
  • Final output can be upscaled to 1080p or 4K, aimed at freelancers and studios selling finished video work.

Google wants a slice of the money freelancers and small studios are starting to make from AI video. Today it rolled out Gemini Omni 1.1 Flash, a new version of its generative video model, meaning software that turns a written prompt into a moving clip. The price and the plumbing are the real story here.

What actually changed?

Omni 1.1 Flash adds four practical controls aimed at people who sell video for a living. You can extend a scene, lock the first and last frame, draft cheaply in low resolution, and upscale to 4K. There's also a reference-video input: feed the model up to three seconds of existing footage to keep characters looking consistent across shots.

Scene extension is the headline feature. The model reads up to 10 seconds of a clip you already have, then generates the next 10 seconds in the same style, repeatable up to a total of 40 seconds. Earlier versions only looked at the final second, which is why AI videos drifted into nonsense after a few beats.

Keyframe control lets you hand the model a starting image and an ending image, then asks it to fill in the movement between them. It's useful for camera orbits, zoom transitions and looping clips, bread-and-butter work for social media ads. We first covered generative video tools at AI2Day on 15 July 2026; our story on an AI film crew making a full-length action comedy on 13 August showed exactly where these controls still fall short.

How much does it cost to use?

Google's pushing developers toward a cheap-draft, expensive-final workflow. Rough previews render in 360p, a low resolution good enough for checking whether an idea works. Google says these drafts run up to 60% faster and at roughly a third of the cost of 720p.

Once you're happy with the draft, you upscale to 1080p or 4K for the client.

Feature Detail
Draft resolution 360p, up to 60% faster, about 1/3 the cost of 720p
Max scene extension 40 seconds total, in 10-second steps
Context window Previous 10 seconds of video
Final output Up to 4K via upscaling
Reference input Up to 3 seconds of video

Exact per-second pricing sits in Google's developer docs and shifts by region. Check before you quote a client.

Who is actually going to make money from this?

The honest answer: people already selling video services. Small ad shops, wedding editors, product marketers, YouTubers with a real audience. The tools cut hours off a job they already know how to price.

Beware the survivorship stories. Plenty of people are posting free clips into a saturated feed and earning nothing. The tool doesn't bring the client.

Freelancers quietly winning tend to share two things: an existing customer list and a niche. AI video makes those jobs faster to deliver, not easier to find.

Where can you try it?

Google says Omni 1.1 Flash is live in Google AI Studio for developers, on the Gemini Enterprise Agent Platform for larger customers, and inside Google Flow for paying Google AI Plus, Pro and Ultra subscribers worldwide. Scene extension is also in the Gemini app for those same paying tiers.

Free users get to look through the window. Paying users get the tools.

Before subscribing to anything, list three real clients or projects where a 40-second AI clip would earn or save you money this month. If that list is empty, the free tier is enough to learn on. Start paying once it isn't.

© 2026 AI2Day