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

Open original

Google has made Gemini Omni 1.2 Flash generally available, a video generation model that now produces 15-second clips and costs 30% less than the previous version. It adds multi-turn editing, tighter consistency across shots, and more natural speech and camera control.

  • Price drops 30% from Omni 1.1 to $0.07 per second for 720p video, with a draft-at-low-res, upscale-later workflow to keep costs down.
  • Clips can now run up to 15 seconds, up from 10, with duration set in one-second steps from 3 to 15.
  • You can refine a clip turn by turn (swap backgrounds, change lighting, switch aspect ratio) without regenerating the whole thing.
  • Character, product, and brand details like faces, garments, and packaging typography are meant to stay consistent across angles and shots.

Google's Gemini Omni 1.2 Flash is now generally available through the Gemini API and Gemini for business. The headline changes are a 30% price cut from Omni 1.1, down to $0.07 per second for 720p video, and clips that run up to 15 seconds, up from 10. Duration can be set in one-second steps between 3 and 15 seconds. Google suggests a cheaper workflow: draft at 360p or 720p, fix small problems with a follow-up edit rather than regenerating the whole clip, then upscale the keepers to 1080p or 4K. The bigger shifts are in consistency and editing. Users can feed in character, product, or style images, short motion clips up to three seconds, and audio cues, and the model is supposed to hold faces, garments, packaging text, and brand colors steady across camera angles and cuts. Multi-turn editing lets you change a background, swap a product, adjust lighting or weather, or change the aspect ratio on an existing clip without starting over. Google's examples include a virtual try-on sequence built from one model photo and front-and-back shots of three outfits, and a single clip whose background gets swapped four times while the characters' performances stay intact. Dialogue, lip-sync, and multi-speaker handling are also improved, and camera work is smoother, with fewer unwanted jump cuts on FPV flythroughs and orbits. Availability varies by surface. Developers can prototype in Google AI Studio and check the docs and pricing. Enterprises can try it in the Google Cloud console, and Provisioned Throughput is offered for steady workloads. Consumers get it inside Flow, Google Vids (15-second clips at no extra cost for Google accounts and Workspace subscribers), Google Ads, and the Gemini app, where 15-second video is rolling out to Google AI Ultra subscribers first. Every output carries an imperceptible SynthID watermark. Google says fast sports action, rapid script changes, and complex physics are still works in progress, which is worth keeping in mind if your use case depends on those.

Ask about this article