Google's Gemini Omni 1.2 Flash Makes 15-Second AI Video

The feature Google is pushing hardest is what it calls conversational editing. Instead of regenerating a whole clip to fix one thing, users can change the background, lighting or aspect ratio step by step in an ongoing conversation, and Google says this works across the full 15-second length.

Google also claims stronger reference adherence. The model accepts character images, product photos, motion clips of up to three seconds and audio cues, and is supposed to keep faces, clothing, packaging typography and brand colors consistent across different camera angles. One example shows a virtual try-on clip generated from a single model photo and product shots of three outfits, producing a 15-second sequence with 360-degree turns.

The company says overall visual consistency has improved too, with less drift or morphing across motion and camera transitions. Dialogue is described as more fluid and natural-sounding, with clearer lip-sync and the ability to tell multiple speakers apart. Complex moves such as FPV drone flights or dramatic zoom-outs are said to arrive with significantly fewer unwanted jump cuts.

For anyone in Tunisia making ads, social content or product videos, this lowers two barriers at once: clip length and cost. Fifteen seconds is enough for a real spot rather than a fragment, and $0.07 per second at 720p — roughly 0.20 TND — puts short drafts within reach of freelancers and small studios, especially using the low-resolution-first workflow.

There is a catch worth knowing. Full access through the Gemini app is tied to the Google AI Ultra tier, while the Gemini API, Google AI Studio and Flow are the practical routes for most professionals. Google has not announced any Tunisia-specific pricing or availability, so treat local access as unconfirmed for now.