Gemini Omni 1.2 Flash: 15-Second Video at $0.07 a Second

Google has made Gemini Omni 1.2 Flash generally available for the first time, and the release tackles the two complaints that follow every AI video model: clips are too short, and small changes mean starting over. The new version generates clips of up to 15 seconds natively and lets users refine them through a back-and-forth conversation. It also costs 30% less than its predecessor.

• Gemini Omni 1.2 Flash generates clips of 3 to 15 seconds, up from a 10-second limit on Omni 1.1
• The price falls 30% to $0.07 per second for 720p video, or about $1.05 for a full 15-second clip by our math
• Multi-turn editing can swap a product, background, lighting or aspect ratio without re-generating the clip
• The model is generally available on the Gemini API and Google Cloud, and outputs carry an imperceptible SynthID watermark

What Changed in Gemini Omni 1.2 Flash?

Duration is the easiest upgrade to measure. Users can set clip length in one-second steps from 3 to 15 seconds, which Google says is enough for a product explainer or a complete social ad in a single generation.

The more significant technical claim concerns consistency. Generative video has struggled with drift, where a face, garment or label slowly morphs between frames. Google says 1.2 Flash holds character, object and scene details more firmly from the first frame to the last, and across camera angles. Users can guide generation with character, product or style images, motion clips of up to three seconds, and audio cues.

Google also reports better lip-sync and dialogue with distinct voices for multiple speakers, plus steadier continuous camera moves with fewer unwanted jump cuts. These are the company's own assessments, drawn from its demonstration clips, so independent testing will show how well they hold across everyday prompts.

How Much Does Gemini Omni 1.2 Flash Cost?

At $0.07 per second for 720p, the headline price is low, but the suggested workflow matters as much. Google recommends drafting at 360p or 720p, fixing details with a follow-up edit instead of generating a new clip, and only then upscaling the best result to 1080p or 4K. That approach treats video more like an iterative document than a one-shot render, and it reduces the number of expensive attempts per finished clip.

It is a notable step from the consumer-focused debut of Omni in Google's June Pixel Drop, and it adds competition among multimodal video models such as MiniMax H3.

Where Can You Use Gemini Omni 1.2 Flash?

The rollout spans several Google products:

• Developers can prototype in Google AI Studio and build through the Gemini API
• Businesses can use the Google Cloud console, with Provisioned Throughput available
• Everyday users can generate 15-second clips at no additional cost in Google Vids, and direct scenes in Flow
• Gemini app access to 15-second video is rolling out to Google AI Ultra subscribers

Google is candid about the remaining limits, saying work continues on fast sports action, rapid script changes and complex physics. That is a helpful guide for anyone choosing what to try first: product shots and character-driven scenes look like the model's strongest territory today. For more on generative models, see our AI coverage.

Sources: Google — November 5, 2026; Google DeepMind — November 5, 2026.