Gemini Omni 1.2 Flash Stretches Native Clips to 15 Seconds and Cuts 720p Video to $0.07 a Second

Google has made generally available Gemini Omni 1.2 Flash, extending native video generation and editing to fifteen seconds, tightening reference adherence across camera moves, and cutting list price by about 30% to $0.07 per second of 720p video on the Gemini API and Google Cloud. The November 5, 2026 Google DeepMind developer blog by product manager Anish Nangia positions 1.2 Flash as the production path for businesses and developers who need longer social spots, product explainers, and multi-turn ad variants without regenerating every take from scratch. Duration can be set in one-second steps from 3 to 15 seconds, up from a 10-second ceiling on Omni 1.1 Flash. Outputs carry an imperceptible SynthID watermark; Provisioned Throughput is available for enterprise capacity planning.

The workflow Google emphasizes is cost as much as length. Teams can draft at 360p or 720p, fix a detail with a follow-up conversational edit instead of burning a full re-render, then upscale keepers to 1080p or 4K. At the new 720p rate, a full 15-second clip works out to about $1.05 before edits. Google Vids users can generate 15-second clips at no additional cost under the account/Workspace paths the post lists, while the Gemini app path for 15-second video is described as rolling out to Google AI Ultra subscribers.

References, consistency, and edits without a fresh brief

Reference adherence is the other headline. Creators can combine character, product, or style stills with short motion clips up to three seconds and audio cues; Google says faces, garments, packaging typography, and brand colors hold more reliably across angles—aimed at virtual try-on, catalog spins, and branded ads built from existing photography. Consistency claims cover less morphing when interpolating first-to-last keyframes or extending a clip into a longer sequence. Multi-turn conversational editing lets a team swap backgrounds, weather, lighting, or aspect ratio on the same 15-second take rather than re-briefing the whole generation. Speech improvements called out include more natural dialogue, clearer lip-sync, and distinct multi-speaker conditioning for Flow, Gemini for business explainers, and Google Vids avatar/narration stacks. Camera control examples stress continuous FPV flythroughs, zoom-outs, and orbits with fewer unwanted jump cuts.

Google is explicit about remaining weak spots: fast sports action, rapid script changes, and complex physics still need work. That caveat keeps the release from reading as a universal video model; it is positioned for scaled marketing, product, and creative pipelines where reference lock and cheap iteration matter more than chaotic athletic motion.

Where it ships and what “GA” changes

Availability spans the developer and business surfaces Google usually splits. Developers prototype in Google AI Studio and call the Gemini API; businesses use Google Cloud console paths with enterprise infrastructure and security framing. Creative tools called out include Flow for directed scenes and multi-turn edits, Google Vids for 15-second generations, the Gemini app (Ultra rollout), and Google Ads for campaign-scale video. “Generally available” here means the model is no longer a limited preview for production billing and support expectations on API and Cloud—not that weights are downloadable. Omni 1.2 Flash remains a hosted multimodal video system: text, images, short video, and audio in; video with audio out.

The concrete changes are the duration bump, the published $0.07/s 720p price, multi-turn edit economics, and SynthID marking—not third-party Elo boards Google does not publish for 1.2. Compared with same-week open-weight model drops, this is closed infrastructure with a clear unit price and a workflow that treats editing as conversation rather than regeneration. Teams already generating Omni 1.1 spots get longer native clips and a lower per-second bill; teams evaluating video APIs get a concrete cost anchor and an honest list of motion regimes Google still flags as unfinished.

Gemini Omni 1.2 Flash is therefore a production video update more than a research surprise: fifteen seconds natively, cheaper seconds on the meter, stronger hold on brand references, and conversational edits that keep the same clip alive across audience variants—shipped GA on the Gemini API and Google Cloud on November 5, 2026.