Google Makes Gemini Omni 1.2 Flash Generally Available: 15-Second Videos at $0.07 Per Second, 30 Percent Cheaper

Google announced Gemini Omni 1.2 Flash on Thursday, an update to its video generation model that stretches clips from 10 to 15 seconds and cuts the price by roughly 30 percent. More significant than either number: this is the first time the Gemini Omni line is generally available, meaning companies can build it into production tools rather than testing behind a preview flag.

The headline price for developers is $0.07 per second of 720p video, 30 percent below Omni 1.1. Clip length is now set in one-second steps between 3 and 15 seconds, and Google's cost guidance is to draft ideas at 360p or 720p first, then upscale only the keepers to 1080p or 4K.

The model's editing story is conversational rather than regenerate-and-pray. A developer can ask to swap a product, change the weather or lighting, or switch the aspect ratio across a full 15-second clip one request at a time, and the model carries the rest of the scene forward instead of re-rolling it. Google says reference-following also improved: inputs can include character, product or style images, audio hints, and motion clips up to three seconds, with faces, clothing, packaging text and brand colors meant to stay consistent from shot to shot.

Dialogue gets more natural lip-sync and distinct voices for multiple speakers, and long camera moves such as FPV flythroughs and orbits should produce fewer unwanted jump cuts. Every output carries Google's invisible SynthID watermark.

Distribution is broad. The model is available through the Gemini API and Google Cloud, with Provisioned Throughput for heavier workloads, and it lands in Google Flow, Google Ads and Google Vids — Vids users get the 15-second clips at no extra cost. In the Gemini app, 15-second generation starts with Google AI Ultra subscribers as the rollout continues.

Google's demo clips sketch the intended use cases: a 15-second virtual try-on built from one model photo and front-and-back shots of three outfits, a single unbroken FPV flight through a plaza and up into the Pantheon's dome, and a macro science explainer where spelled-out labels follow the motion. Google itself flags the weak spots — fast sports action, complicated physics and scripts that change quickly — and notes that its demo clips are curated, so real-world quality will only show as developers push the model.

The release follows Google DeepMind's Genie 4 world model in late October and tightens the pricing pressure on the AI video market, where Kuaishou's Kling, Runway, OpenAI's Sora line and a crowd of Chinese startups are all competing on cost per second. At 7 cents per second of 720p, Google is betting that volume production — ads, product videos, social content — matters more than premium cinematography.

For the industry, the GA label is the real signal. Video models have spent the past year accumulating impressive demo reels behind limited previews; Gemini Omni 1.2 Flash is the moment one of the major labs put a stable price, an SLA-backed cloud endpoint and a workspace footprint behind one.