Google Gemini Omni 1.2 Flash Launches With 15-Second Video

Consistency and editing

Users can steer generation with images of people, products or styles, short motion clips of up to three seconds, and audio cues. Google says faces, clothing, packaging text and brand colors stay consistent across a clip even when the camera angle changes. That allows virtual try-ons, rotating catalog views and brand videos built from existing product photos or character sheets.

Editing works as a conversation: backgrounds, lighting, weather and aspect ratio can be changed without starting over, which makes it easier to produce ad variants for different audiences and channels. Dialogue between several speakers sounds more natural, with clearer lip sync, and continuous camera moves such as drone-style FPV shots contain fewer unintended cuts. Fast sports action, abrupt script changes and complex physics are still being improved.

Availability

The model runs on Google Cloud's enterprise platform and supports Provisioned Throughput. Output carries SynthID, an invisible digital watermark. Adobe Firefly, Figma Weave and Magnific already use it, and early testers include Monks and Whering, which tried it for virtual try-on.

Developers can start in Google AI Studio, and businesses through the Google Cloud console. In Google's own products it is available in Flow, Google Vids, the Gemini app (rolling out to Google AI Ultra subscribers) and Google Ads. Google Vids can generate 15-second clips at no extra cost.