Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips

Gemini Omni 1.2 Flash, Google's latest model for generating and editing video, produces longer clips for less money: up to 15 seconds per generation, up from 10 with Omni 1.1, at $0.07 per second of 720p video, which Google says is 30% below Omni 1.1 (Google DeepMind). Clip length can be set in one-second steps from 3 to 15 seconds. It is also the first time Omni is generally available, through the Gemini API and Google Cloud, where businesses can reserve capacity with Provisioned Throughput.

Google's cost argument also rests on not regenerating whole clips. A user can draft cheaply at 360p or 720p, then refine the same clip turn by turn, for example swapping a product or background or changing the lighting, and upscale the chosen version to 1080p or 4K. Low-resolution drafting and 4K upscaling arrived with Omni 1.1. The turn-by-turn editing is new in 1.2, and Google says it works across full 15-second clips. Generation can be guided by images of characters, products or styles, short motion clips of up to three seconds, and audio cues.

The quality claims are Google's own, shown through examples it selected, including comparisons with Omni 1.1, rather than measured results. Google says faces, clothing, packaging text and brand colors stay consistent across camera angles, with less drift between the first and last frame, clearer lip-sync and fewer unwanted jump cuts. It says work continues on fast sports action, rapid script changes and complex physics. The model is also rolling out in Google Vids at no extra cost and in the Gemini app for Google AI Ultra subscribers, and every output carries SynthID, a watermark Google describes as imperceptible.