Google has unveiled Gemini Omni 1.2 Flash, the latest iteration of its video generation and editing model, now generally available across the Gemini API and Gemini for business. The headline features: native clips up to 15 seconds, multi-turn conversational editing, and a 30% price reduction compared to Omni 1.1. According to Google DeepMind product manager Anish Nangia, the model is designed to make video creation faster and more cost-effective at scale.
- Longer clips: Generate and edit videos up to 15 seconds, up from 10 seconds, with duration adjustable in one-second increments from 3 to 15 seconds.
- Conversational editing: Refine a clip turn by turn—swap products, backgrounds, lighting, or aspect ratios—without regenerating from scratch.
- Lower cost: Priced at $0.07 per second for 720p video, 30% less than Omni 1.1, with similar savings at other resolutions.
What’s new in Omni 1.2 Flash
Beyond the extra five seconds, the model introduces a more intuitive workflow. Users can supply character, product, or style reference images, short motion clips of up to three seconds, and audio cues. Google claims the model maintains consistent faces, garments, packaging typography, and brand colors across camera angles, reducing visual drift between the first and last frame. This makes it suitable for virtual try-on sequences, catalog spins, and branded video ads.
Editing is where the model truly shines. Instead of re-prompting from zero, creators can issue follow-up instructions to change a product, background, lighting, weather, or aspect ratio across the full 15-second clip. This multi-turn conversational editing works across the entire duration, enabling rapid production of ad variants for different audiences and placements.
“Faces, garments, packaging typography, and brand colors stay consistent across camera angles,” Nangia wrote in the announcement.
Speech generation has also improved, with more fluid dialogue, clearer lip-sync, and distinct multi-speaker conditioning. However, Google acknowledges that fast sports action, rapid script changes, and complex physics are “still in development.”
Pricing and availability
The price cut is a strategic move to encourage volume adoption. At $0.07 per second for 720p, a full 15-second clip costs about $1.04. Drafts can be generated at 360p for roughly $0.35 per 15 seconds, and final outputs can be upscaled to 1080p or 4K. This tiered workflow lets teams iterate cheaply before committing to higher-resolution renders.
Gemini Omni 1.2 Flash is available through the Gemini API via Google AI Studio, Google Cloud for enterprises, and the Flow tool for scene direction. It is also integrated into Google Vids—where 15-second clips are free with a Google or Workspace account—the Gemini app (initially for Google AI Ultra subscribers), and Google Ads for video campaigns. This broad integration distinguishes Google from standalone generators like ByteDance’s Seedance 2.5, which offers 4K video up to 30 seconds. Google is not just selling a model; it is embedding video generation into the tools people already use.
Watermarking and safety
All outputs carry an imperceptible SynthID watermark by default, aligning with Google’s policy on labeling AI-generated content. While visible watermarks in Gemini have become optional, the invisible SynthID remains always active. This is a thin patch for a growing problem: as credible video generation becomes as cheap as a cup of coffee, the internet will fill with material indistinguishable from real footage, and an invisible watermark that few check may not be enough.
Early customers include Artlist, FLORA, invideo, Krea, Picsart, Photoroom, and QuickFrame AI, all using the model for short, commercial, repeatable jobs. Photoroom’s CEO Matt Rouif noted that version 1.2 brings 15-second videos, a top user request, and that general availability with SLAs lets them rely on it in production. For businesses producing dozens of ad pieces per week, the combination of price and integration into Vids, Ads, and Workspace may outweigh raw quality. The Flash version, rather than any flagship model, is what truly moves the market.