Google's Video AI Gets Longer Clips and a Lower Price

Google has made Gemini Omni 1.2 Flash generally available, moving its fast, low-cost video model out of preview and into the Gemini API and Gemini for business. Developers can prototype with it in AI Studio. Google announced the rollout on November 5.

Two changes matter most. Clips can now run up to 15 seconds, up from 10, and the length is adjustable from 3 to 15 seconds in one-second steps. The price has come down by about 30 percent, to roughly seven cents per second for 720p video, according to The Decoder. Google also says the model edits more cleanly and makes fewer errors than the version it replaces.

Where it shows up

The model is reaching a long list of Google surfaces: the Gemini API, AI Studio, the Google Cloud Console, the Flow filmmaking tool, Google Vids, the Gemini app, and Google Ads. In the consumer app, the 15-second clips are limited to Google AI Ultra subscribers for now. Google's own model card lays out the technical details.

Every clip the model produces carries a SynthID watermark, an imperceptible marker Google embeds so that a video's AI origin can be detected later. That provenance signal is becoming standard across the major generators, and it is quietly doing a lot of work as synthetic footage spreads.

Cheap enough to be ordinary

The headline figure is the cost. At seven cents a second, a finished 15-second clip runs about a dollar before editing. Pricing like that moves AI video from a novelty into routine production for advertising, marketing and social teams, which is plainly the audience Google is aiming at with placements inside Vids and Ads.

The quality is still short of a trained crew with a camera, and 15 seconds is not a film. But the trend line is clear. Each release makes synthetic video a little longer, a little better and a good deal cheaper, and that trajectory is the real story for anyone who makes or watches video for a living.