Google launched Gemini Omni 1.2 Flash, an updated video model available on the Gemini API, Google Cloud, Google Vids, and Gemini for business. The update extends single clip lengths from 10 seconds to 15 seconds. Users can choose exact durations between 3 and 15 seconds in increments of one second.

The company lowered generation prices by 30 percent compared to version 1.1. Rendering 720p video now costs 0.07 dollars per second. Users can draft sequences in 360p or 720p, refine elements using follow up prompts, and upscale approved clips to 1080p or 4K resolution.

The new model adds multi turn editing, allowing creators to adjust scenes conversationally. Prompts can swap backgrounds, alter lighting, or replace items while keeping actor identity, camera paths, and garment details intact. It also improves dialogue sync with distinct voice conditioning across multiple speakers.

Why it matters

For video creators and marketing teams, longer clips and conversational editing cut production iteration time. Instead of spending credits to generate a completely new video for a minor fix, creators can simply ask the model to swap a prop or alter a setting.

Stronger visual adherence also helps ecommerce sellers turn still product shots into clean rotating previews and virtual try on reels without brand assets warping or drifting between frames.

Try it yourself

Use this prompt to build a multi shot video concept ready for Gemini Omni 1.2 Flash.

Act as an AI video director. Write a detailed generation prompt for Gemini Omni 1.2 Flash to produce a [duration between 3 and 15] second video showcasing [product or subject]. Include clear instructions for continuous camera movement such as [drone flythrough or slow orbit], specify the background setting of [desired environment], and list consistent visual references for [brand colors, materials, or clothing details] that must remain stable across all frames.