Google's Gemini Omni 1.2 Flash video model goes GA with 15-second clips

What happened

Clip length can be set from 3 to 15 seconds, in one-second steps. Omni 1.1 was capped at 10 seconds.

The price is $0.07 per second of 720p video, 30% lower than Omni 1.1 Flash.

Drafts render at 360p or 720p. Only final versions are upscaled to 1080p or 4K, so iterations stay at lower resolution until a clip is chosen.

Conversational editing can change the background, lighting or aspect ratio of an existing clip without regenerating it.

The model is available through the Gemini API, AI Studio, Google Cloud, Flow, Google Vids and Google Ads. In the Gemini app, 15-second video is rolling out to AI Ultra subscribers.

Every output carries an invisible SynthID watermark.

Why it matters for creators

A cheaper per-second rate from Google raises price pressure across AI video, and gives creators a useful benchmark when comparing cost per clip.

The draft-then-upscale approach and conversational editing both target the same problem: paying for full regenerations while a shot is still being worked out. Fixing lighting or reframing to a new aspect ratio on an existing clip means a good take does not have to be thrown away.

The SynthID watermark is invisible but present on every output, which matters for anyone publishing to platforms that check for AI content.