Google launched Gemini Omni 1.2 Flash today, making its video generation model generally available with native 15-second clips and a 30% price cut to $0.07 per second of 720p video. The release, announced on Google's blog by Google DeepMind product manager Anish Nangia, adds multi-turn conversational editing and closer adherence to reference images, motion clips and audio cues.
Gemini Omni 1.2 Flash generates clips of up to 15 seconds, up from a 10-second maximum in Omni 1.1. Users can set duration in one-second increments from 3 to 15 seconds, draft at 360p or 720p, and then upscale the best results to 1080p or 4K. Output defaults to 720p in 16:9 landscape, with 9:16 portrait also supported.
The feature aimed at production teams is conversational editing. Instead of regenerating a clip from scratch, a user can ask the model to swap a product or background, change the lighting or weather, or switch the aspect ratio, and the model leaves unmentioned elements of the scene intact. Google said the model can combine character, product and style images, motion clips of up to three seconds, and audio cues as references.
Key facts
- 01Google made Gemini Omni 1.2 Flash generally available on November 5, 2026, with native clips of up to 15 seconds, up from 10.
- 02720p video costs $0.07 per second, 30% less than Gemini Omni 1.1 Flash, with similar savings at other resolutions.
- 03Drafts render at 360p or 720p and can be upscaled to 1080p or 4K.
- 04Gemini API requests to Omni 1.1 Flash will route automatically to 1.2 on December 3, 2026.
- 05Picsart, Photoroom, invideo, Krea and QuickFrame AI are among the early customers Google Cloud named.
Google is pitching the release at the video businesses produce in volume. In a Google Cloud blog post, senior product manager John Nack described the model as built for everyday business video "like product videos, virtual try-ons, and ads," and said it follows image, video and audio references more closely than before.
“We chose Gemini Omni Flash to power our Video Ad Maker and Product Animator because it balances quality, speed and cost.”— Matt Rouif, CEO and Co-founder, Photoroom
On price, Google said 720p output costs $0.07 per second, 30% less than Gemini Omni 1.1 Flash, with similar savings at other resolutions. On the paid Gemini API that works out to roughly $0.70 for a 10-second clip and $1.04 for a 15-second clip, and the model is not offered on the API's free tier. Gemini API requests to Omni 1.1 Flash will route automatically to 1.2 on December 3, 2026, at the lower rates and without code changes.
Gemini Omni 1.2 Flash is available through the Gemini API in Google AI Studio, the Google Cloud console and enterprise API with Provisioned Throughput, Flow, Google Ads and Google Vids, where 15-second clips come at no additional cost with a Google account or Workspace subscription. In the Gemini app, 15-second video creation is rolling out to Google AI Ultra subscribers. Audio references on Google Cloud are gated behind allowlisting.
Google Cloud named Artlist, FLORA, invideo, Krea, Picsart, Photoroom, QuickFrame AI and Runway among the companies building on the model. Photoroom uses it to power its Video Ad Maker and Product Animator, while QuickFrame AI said 15-second generations at 1080p let users create a complete connected-TV ad in a single generation.
“Now that it is generally available, we can bring this quality to over 100M Picsart creators.”— Narek Hayrapetyan, Head of Product and Partnerships, Picsart
Google DeepMind's side-by-side evaluations put Omni 1.2 ahead of Omni 1.1 with a 62% win rate on text-to-video visual consistency and a 63% win rate on multi-reference identity consistency. A separate video extension feature appends 3 to 10 seconds per turn, up to 30 seconds in total, using the final 10 seconds of a clip as context.
“If you need fast video at scale, it's one of the best we've tested.”— Shivam Rastogi, VP of Engineering, invideo
Google said development is ongoing to improve fast sports action, rapid script changes and complex physics. In the current Gemini API, English is fully supported while other languages have not been evaluated, and voice editing and uploaded audio inputs are not yet supported.
Every clip the model produces carries an imperceptible SynthID watermark, which Google said can be detected programmatically for provenance verification.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
