---
title: "Google releases Gemini Omni 1.2 Flash with longer video clips and lower pricing"
description: "Google releases Gemini Omni 1.2 Flash with 15-second clips. Explore paid API pricing, audio-reference restrictions, rollout details and benchmark limitations."
url: "https://promptszone.com/google-gemini-omni-1-2-flash-video-release-pricing-access"
published: "2026-11-05T21:00:06Z"
categories: ["New Models"]
---

# Google releases Gemini Omni 1.2 Flash with longer video clips and lower pricing

Gemini Omni 1.2 Flash adds 15-second video and lower output pricing. Access restrictions, token billing and evaluation limits matter before adoption.

![Original diagram of Gemini Omni 1.2 Flash’s draft, refine and upscale workflow, with access caveats and the official Gemini family logo.](https://promptszone.com/assets/uploads/d65a4b036084a92b23fe0e926391ed4f.png)

**Summary:**

Google’s video model is generally available on developer platforms. The release expands clip length, but audio-reference access, app rollout and workflow limits still need checking.

Google has released Gemini Omni 1.2 Flash, expanding its video-generation model to 15-second clips and lowering the advertised cost of 720p output. The November 5 launch brings general availability on developer platforms, while consumer-app access still depends on the product and rollout.

The practical question is how much usable footage a team can produce after revisions. A cheaper generated second helps, but consistency, editing reliability and the number of rejected takes still determine the cost of a finished video.

## Longer clips and a more iterative workflow

Google’s [release announcement](https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/) describes native generation and editing up to 15 seconds, compared with 10 seconds for Omni 1.1. It promotes drafting at 360p or 720p, refining a clip through follow-up instructions, then upscaling selected results to 1080p or 4K. Better reference adherence, speech and visual continuity are Google’s claims, not independently reproduced results in this report.

Google also acknowledges ongoing work on fast sports, rapid script changes and complex physics. Those are useful stress tests for prospective users, rather than details to overlook in a polished demonstration.

## General availability has product-specific limits

The [Google Cloud launch post](https://cloud.google.com/blog/products/ai-machine-learning/gemini-omni-1-2-flash-generally-available) confirms general availability, Provisioned Throughput and SynthID watermarking. It also includes an important qualification: audio references require allowlisting. A model’s ability to accept a modality does not establish that every customer can use it immediately.

The broader announcement lists Flow, Google Vids and developer tools among the access routes. In the Gemini app, the new 15-second experience is rolling out to Google AI Ultra subscribers. Check the relevant product’s availability and account limits before planning a production workflow.

The [Gemini API changelog](https://ai.google.dev/gemini-api/docs/changelog) adds distinctions that a headline duration can hide: first-and-last-frame interpolation runs from 3 to 10 seconds; extensions add 3 to 10 seconds per turn, up to a total length of 30 seconds. It lists support for up to six reference images and doubled paid-tier rate limits. Requests to gemini-omni-1.1-flash are scheduled to route automatically to 1.2, at the new rates, from December 3, 2026.

## What the price means

The [standard Gemini API price list](https://ai.google.dev/gemini-api/docs/pricing#gemini-omni-1.2-flash) shows no free tier for Omni 1.2 Flash. Rates are US$1.50 per million input tokens for text, image or video, US$9 per million text-output tokens, and US$12 per million video-output tokens.

At Google’s stated 5,792 output tokens per second of 720p video, the video component is approximately US$0.07 a second, or US$1.04 for 15 seconds. This is not an all-inclusive project quote: input, any text output, additional generations and higher-resolution output can add cost. Google describes the video-price reduction from 1.1 as 30%.

## Why there is no model leaderboard here

The launch materials reviewed do not establish a matched, multi-benchmark comparison against leading video models. Demonstrations and vendor claims cannot supply missing scores or consistent test conditions. The [November model card](https://deepmind.google/models/model-cards/gemini-omni-1-2-flash/) also refers readers to the earlier Omni Flash card for several technical and evaluation details. This article therefore does not present a numerical quality ranking.

For an evaluation, use the same rights-cleared references and instructions across candidate models. Record usable-take rate, identity and product-detail drift, lip-sync errors, edit success, latency and total billed cost. Review a longer sequence frame by frame, and check whether a successful edit introduces a new defect elsewhere. These are suggested checks, not tests performed for this report.

Visual: original explanatory diagram, not model output. Official Gemini family-logo fallback used unchanged apart from proportional sizing. Logo source: [Google image library](https://blog.google/image-library/), which permits publication with credit. Source: Google. No endorsement implied.