11-19 更新 实时
首页/ AI 大模型/ 正文

Gemini Omni 1.2 Flash:快速低成本,支持更长片段

AI 大模型 来源:Google Blog 2026-11-06 01:00 0 阅读 机器翻译
Gemini Omni 1.2 Flash:快速低成本,支持更长片段

产品经理,Google DeepMind

您的浏览器不支持音频元素。

我们正在推出 Gemini Omni 1.2 Flash,它支持原生 15 秒视频生成、更强的参考一致性,以及多轮对话式编辑。本次更新将成本降低了 30%,并为您专业项目带来更好的一致性。从今天起,您可以在 Gemini API、Google Vids 和 Gemini for business 中使用这些能力,大规模制作高质量视频内容。

摘要由 Google AI 生成。生成式 AI 仍处于实验阶段。

我们正式推出 Gemini Omni 1.2 Flash,支持 15 秒视频并提升了参考一致性。

现在你可以通过多轮对话编辑片段,无需从头开始即可更快完成调整。

我们将价格降低了 30%,帮助你以更低成本制作高质量视频内容。

该模型已在 Gemini API 和 Google Cloud 上正式可用,满足你的生产需求。

摘要由 Google AI 生成。生成式 AI 仍处于实验阶段。

今天,我们正式推出 Gemini Omni 1.2 Flash,并在 Gemini API 和 Gemini for business 中全面开放使用。Gemini Omni 1.2 Flash 引入了原生 15 秒视频生成、多轮对话式编辑能力,以及增强的参考遵循能力。我们已将价格较 Gemini Omni 1.1 下调 30%,720p 视频定价为每秒 0.07 美元。

以下是新功能的概览:

大规模下快速且高性价比

除了更低的价格,整个工作流也能持续压缩成本:你可以先用 360p 或 720p 快速生成草稿,通过后续编辑修正某个细节,而不必重新生成整段视频,然后把最满意的作品超分至 1080p 或 4K。

更长的片段,最长可达 15 秒

现在你可以原生生成和编辑最长 15 秒的视频,相比 Omni 1.1 的 10 秒有所提升。你还可以按 1 秒为增量,在 3 到 15 秒之间指定时长,让你有足够空间在一次生成中讲述完整的故事、制作产品讲解视频,或产出完整的社交广告片。

更好的参考遵循能力

你可以结合人物、产品或风格图像、短动作片段(最长三秒)以及音频线索来引导生成——而 Omni 1.2 Flash 会从头到尾保留住这些细节。人脸、服装、包装字体和品牌配色在不同机位下保持一致,从而轻松将现有的产品摄影或角色设定图转化为虚拟试穿片段、商品目录旋转展示和品牌视频广告。

将一张模特照片与三套服装的正反面产品图相结合,Gemini Omni 1.2 Flash 可生成一段 15 秒的虚拟试穿序列,涵盖 360 度转身和匹配剪辑,同时保持模特的身份特征以及每件服装的剪裁、垂坠感和背面细节一致。

你可以让角色、物体和场景从首帧到末帧保持更强的一致性,在运动和镜头转场中减少视觉漂移或形变。无论是在首尾关键帧之间进行插值,还是延长片段以构建更长的序列,产品、角色、道具和环境都能在逐个镜头之间保持一致。

在给定同一张静态影棚布光的展览海报和五层动画提示词时,Gemini Omni 1.1 Flash 对参考图像的还原程度不如 1.2。Gemini Omni 1.2 Flash 在灰粉色环形天幕上沿着升起的六边形基座逐层搭建海报,依次拼出带棱纹的玻璃、温暖的灯丝、多面水晶和盘绕电线的字母造型 HALO、GLOW 与 LUMEN,并让每一种材质、阴影和字体细节一直到最后定格都保持一致。

在对话中更轻松地编辑

你可以逐轮精修同一个片段——例如替换产品或背景、更改光照或天气,或切换画幅比例——而不必从头重做。多轮对话式编辑可作用于完整的 15 秒片段,让你无需重新撰写需求说明或重新生成,就能轻松为不同受众和投放位置制作广告变体。

从一个已生成的片段出发,多轮对话式提示词逐轮替换背景——从条纹凉亭帐篷,到橡树林、有白马的马厩、集市,再到石砌城堡——同时保持两个角色的表演、盔甲反光和镜头取景不变。

你可以生成对话更流畅、更自然,口型同步更清晰,并具备明确多说话人条件控制的场景。无论你是在 Flow 中执导多角色对话,在 Gemini for business 中制作视频讲解,还是在 Google Vids 中用虚拟形象、音乐和旁白制作精致的专业视频,这一点都会有所体现。

以一张温室场景图像,外加四个角色、口型和一秒语音参考为条件,Gemini Omni 1.2 Flash 将每个采样语音对应到正确的角色,在 15 秒的动画场景中实现富有表现力的口型同步和自然的喜剧节奏。

更可控的电影质感

你可以执导复杂、连续的镜头运动,比如大幅度的 FPV 穿越飞行,乃至戏剧性的拉远和环绕镜头,多余的跳切显著减少,画面尺度更稳定,观感更精致、更具电影质感。

在两幅建筑参考照片的引导下,Gemini Omni 1.2 Flash 执导了一段连续的 FPV 无人机飞行:穿越广场、穿过青铜门,再沿万神殿的方格穹顶上升至圆洞,全程零跳切。

更长的片段,更流畅的运动,更紧密的一致性

Gemini Omni 1.2 Flash 可生成最长 15 秒的片段,对参考元素的保持能力比以往更强。你可以滚动查看,角色、文字和镜头运动在更长的镜头、多轮生成和转场中始终保持一致,且没有剪辑痕迹。

15 秒长片段:把一个风格帧转化为 4 个镜头、15 秒的宏观科学解说片,其中的标注拼写准确,并能随运动跟踪。

多轮运动迁移:在第一轮生成一只起伏波动的章鱼,然后在第二轮将其替换为一位舞者,其礼服复现了完全相同的流体运动。

首帧到末帧一致性(1.1 与 1.2 对比):在 1.2 Flash 中,把室外起始帧和室内结束帧连接成一次连续的镜头推进。你可以看到屋顶上的猫、汤勺变成筷子等细微差别,以及其他一些小的变化。

不间断的急速拉远:在一个 10 秒的镜头中,从紧凑的海岸线特写连续拉远到岛屿的空中全景,缩放自然、零剪辑。

真实世界的成果,从创意工作者到全球品牌

企业、初创公司和创意工作者已经在各自的工作流中使用 Gemini Omni 1.2 Flash。Gemini Omni 也首次正式全面可用,并由 Google Cloud 的企业级基础设施和安全能力提供支撑。Provisioned Throughput 现已可用,输出内容带有不可感知的 SynthID 水印。

开始使用 Gemini Omni 1.2 Flash

目前相关开发仍在持续推进,以进一步改进快速运动画面、脚本快速变更和复杂物理效果等方面,而 Gemini Omni 1.2 Flash 非常适合需要规模化制作视频的企业、开发者和创作者。你现在就可以开始使用 Gemini Omni 1.2 Flash,它正陆续登陆 Google 的各项产品和开发者平台:

对于使用 Gemini API 的开发者:在 Google AI Studio 中构建原型,体验交互式入门应用,并查看开发者文档和定价。

对于企业:可在 Google Cloud 控制台中试用,并使用该 API 构建企业级解决方案。

对于使用 Google 应用的所有用户:在 Flow 中调度场景并进行多轮编辑;使用 Google 账号或 Google Workspace 订阅,即可在 Google Vids 中免费生成 15 秒片段;在 Gemini 应用中创建 15 秒视频(正陆续向 Google AI Ultra 订阅用户推出);并在 Google Ads 中规模化开展视频广告活动。

译文由英文原文机器翻译,术语按行业惯例统一,可在文末展开原文核对。

查看英文原文(Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips)

Product Manager, Google DeepMind

Text "Gemini Introducing Gemini Omni 1.2 Flash" surroudned by images of a beehive, animals in conversation, a squid, and a woman walking down a street

Your browser does not support the audio element.

We’re rolling out Gemini Omni 1.2 Flash, which offers native 15-second video creation, improved reference adherence, and multi-turn conversational editing. This update lowers costs by 30% and provides better consistency for your professional projects. You can start using these capabilities today across the Gemini API, Google Vids, and Gemini for business to produce high-quality video content at scale.

Summaries were generated by Google AI. Generative AI is experimental.

We are launching Gemini Omni 1.2 Flash with support for fifteen-second videos and improved reference adherence.

You can now edit clips through multi-turn conversations, which allows for faster adjustments without starting over.

We reduced the price by 30% to help you produce high-quality video content at a lower cost.

This model is generally available on the Gemini API and Google Cloud for your production needs.

Summaries were generated by Google AI. Generative AI is experimental.

Today, we’re rolling out Gemini Omni 1.2 Flash, and making it generally available across the Gemini API and Gemini for business. Gemini Omni 1.2 Flash introduces native 15-second video creation, multi-turn conversational editing capabilities, and enhanced reference following. We have dropped the price 30% from Gemini Omni 1.1 making it available at $0.07 per second for 720p video.

Here’s an overview of what’s new:

Fast and cost-effective at scale

Beyond the lower price, the workflow keeps costs down: you can draft fast at 360p or 720p, fix a detail with a follow-up edit instead of re-generating the entire clip, and then upscale your best creations to 1080p or 4K.

Longer clips up to 15 seconds

You can now generate and edit videos natively up to 15 seconds long, up from 10 seconds with Omni 1.1. You can also specify duration in one-second increments from 3 to 15 seconds, giving you room to tell complete stories, create product explainers, or produce full social ad spots in a single generation.

Better reference adherence

You can guide generation combining character, product, or style images, short motion clips (up to three seconds), and audio cues — and Omni 1.2 Flash holds onto those details from start to finish. Faces, garments, packaging typography, and brand colors stay consistent across camera angles, making it easy to turn existing product photography or character sheets into virtual try-on clips, catalog spins, and branded video ads.

Combining one model photo with front-and-back product shots of three outfits, Gemini Omni 1.2 Flash generates a 15-second virtual try-on sequence across 360-degree turns and match cuts while keeping the model’s identity and each garment’s cut, drape, and back details consistent.

You can maintain stronger character, object, and scene consistency from the first to the last frame, with less visual drift or morphing across motion and camera transitions. Whether you’re interpolating between a first and last keyframe, or extending a clip to build a longer sequence, products, characters, props, and environments stay consistent from shot to shot.

Given the same static studio lighting exhibition poster and five-layer animation prompt, Gemini Omni 1.1 Flash does not match the reference image as closely as 1.2 does. Gemini Omni 1.2 Flash builds the poster sequentially layer by layer on the dusty-pink cyclorama rising the hexagonal plinth, assembling the ribbed glass, warm filament, faceted crystal, and coiled-cord letterforms for HALO, GLOW, and LUMEN, and locking every material, shadow, and typographic detail consistently through the final frame.

Easier editing in conversation

You can refine the same clip turn by turn — such as swapping a product or background, changing the lighting or weather, or switching the aspect ratio — instead of starting over from scratch. Multi-turn conversational editing works across full 15-second clips, making it easy to produce ad variants for different audiences and placements without re-briefing or re-generating.

Starting from one generated clip, multi-turn conversational prompts swap the background turn by turn — from a striped pavilion tent to an oak forest, a stable with a white horse, a market fair, and a stone castle — while keeping both characters’ performances, armor reflections, and camera framing intact.

You can generate scenes with more fluid, natural-sounding dialogue, clearer lip-sync, and distinct multi-speaker conditioning. This will show up whether you’re directing multi-character dialogue in Flow, creating video explainers in Gemini for business, or creating polished, pro videos with avatars, music and narration in Google Vids.

Conditioned on one conservatory scene image alongside four character, mouth, and one-second voice references, Gemini Omni 1.2 Flash routes each sampled voice to the right character with expressive lip-sync and natural comedic timing across a 15-second animated scene.

More controllable, cinematic look

You can direct complex, continuous camera movements like sweeping FPV flythroughs to dramatic zoom-outs and orbits with significantly fewer unwanted jump cuts, steadier scale, and a more polished, cinematic look.

Guided by two architectural reference photos, Gemini Omni 1.2 Flash directs one continuous FPV drone flight across the piazza, through the bronze doors, and up through the Pantheon’s coffered dome to the oculus with zero jump cuts.

Longer clips with smoother motion and tighter consistency

Gemini Omni 1.2 Flash generates clips up to 15 seconds long that hold references together better than before. You can scroll through to see characters, text, and camera moves stay consistent across longer shots, multiple turns and transitions with no cuts.

15 second long clips: Turns one style frame into a 4-shot, 15-second macro science explainer with accurately spelled, motion-tracked callouts.

Multi-turn motion transfer: Generates a rippling octopus in Turn 1, then replaces it in Turn 2 with a dancer whose gown mirrors the exact fluid motion.

First-to-last frame consistency (1.1 vs. 1.2): Connects an exterior start frame and interior end frame in one continuous camera push in 1.2 Flash. You can see subtle differences like the cats on the roof, the soup ladle turning into chopsticks, and some other minor changes.

Unbroken crash zoom-out: Pulls continuously from a tight shoreline close-up to an aerial island vista in one 10-second shot with natural scale and zero cuts.

Real-world results, from creatives to global brands

Businesses, startups, and creatives are already using Gemini Omni 1.2 Flash in their workflows. And for the first time, Gemini Omni is now generally available, backed by Google Cloud’s enterprise-grade infrastructure and security. Provisioned Throughput is available, and outputs carry an imperceptible SynthID watermark.

Get started with Gemini Omni 1.2 Flash

While development is ongoing to further improve things like fast sports action, rapid script changes and complex physics, Gemini Omni 1.2 Flash is ideal for businesses, developers, and creators producing videos at scale. You can get started with Gemini Omni 1.2 Flash today as it rolls out across Google’s products and developer platforms:

For developers using the Gemini API: Prototype in Google AI Studio, explore interactive starter apps, and check out the developer documentation and pricing.

For businesses: Try it in the Google Cloud console, and build enterprise solutions with the API.

For everyone using Google apps: Direct scenes and multi-turn edits in Flow, generate 15-second clips at no additional cost in Google Vids with a Google account or Google Workspace subscription, create 15-second videos in the Gemini app (rolling out to Google AI Ultra subscribers), and scale video campaigns in Google Ads.

原文为英文,正文由自动抓取并保留原文。

本文由采集器自动同步

原文发布于 Google Blog(Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips),标题、摘要与正文由机器翻译,术语按行业惯例统一,未经人工逐句校订。

阅读原文 ↗