Google Gemini Omni 1.2 Flash extends video to 15 seconds and cuts price by 30%

How editing works

Google calls the new feature “dialogue editing.” Users can change a clip’s background, lighting, or aspect ratio in stages while keeping the rest unchanged. The company says the feature works across the full length of a video, up to 15 seconds.

Google also recommends a multi-step process designed to reduce costs: first create a clip at 360p or 720p, then refine specific details with additional edits, and upscale to 1080p or 4K only for the final version.

The updated model can generate clips from 3 to 15 seconds long; the previous limit was 10 seconds. 720p video now costs $0.07 per second, 30% less than its predecessor.

Consistency in characters and details

Gemini Omni 1.2 Flash accepts images of characters, product photos, short motion videos up to 3 seconds long, and audio signals as inputs. Google says the model does a better job of preserving faces, clothing, packaging text, and brand colors when camera angles change.

In one example from the company, a 15-second clip with 360-degree turns was created from a single photo of a model and photos of three outfits. The clip preserves the model’s appearance, the cut of the clothes, and how they fit and move.

Google says the new version has fewer visual distortions and fewer changes to details during movement and camera transitions. The model is also supposed to produce smoother, more natural-sounding dialogue, synchronize speech with lip movements more accurately, and distinguish between multiple speakers. It also supports complex, continuous camera movements—such as an FPV drone flight or a sudden pullback—with fewer unwanted cuts.

Dataist's take

The update brings together longer clips, editing of individual details, and lower costs. However, claims about preserving appearance and motion quality come from Google; the announcement contains no independent comparison with the previous model.