What Is Google Gemini? What It Does and What It Costs in 2026

Google Gemini is two things: Google's AI assistant, which you can use free at gemini.google.com and in apps on Android, iPhone, Windows and Mac, and the family of AI models from Google DeepMind that powers it. Gemini replaced Bard in February 2024, now stands in for Google Assistant on Android, and works inside Gmail, Docs and Chrome.

If the name keeps showing up in your inbox, your phone and your browser without anyone explaining it, you're not alone. Google has put Gemini in so many places that it's hard to tell what you're actually being offered. This guide covers what Gemini is, where it lives, what it can do for you at work, what each plan costs, and the privacy settings worth changing on day one.

What is Google Gemini, in plain terms?

Gemini is the name Google uses for both its AI models and the assistant built on them. When someone says "I asked Gemini," they mean the app. When a developer says "we use Gemini," they usually mean the models, reached through Google's API.

The models came first. Google announced Gemini 1.0 on December 6, 2023 in three sizes, Ultra, Pro and Nano, and described it as built to be multimodal from the start: it handles text, images, audio, video and code rather than text alone. The same day, Bard began running on a version of Gemini Pro.

The assistant got the name two months later. On February 8, 2024, Google said Bard would "now simply be known as Gemini" and launched a paid tier alongside it. Since then the models have gone through several generations, and the app has picked up voice chat, research, image and video tools and access to your Google apps.

What's the latest Gemini model?

The newest model is Gemini 4 Argon, which Google announced on September 30, 2026. Google says it can write answers of up to a million tokens, far longer than earlier models. Since October 13, 2026, Google AI Pro and Ultra subscribers can use it in the Gemini app.

You don't need to track model names to use the app well. On Google's plan page, the free plan runs on a smaller, faster Flash-Lite model, Google AI Plus adds the Flash model, and Pro and Ultra add the Pro model and Argon. In practice, the free model is fine for quick questions and rewrites. The bigger models earn their keep on long documents, careful reasoning and research.

Where can you use Gemini?

You can use Gemini on the web, in mobile and desktop apps, inside Google Workspace, in Chrome and on Android devices. According to Google, the web app is available in more than 70 languages and over 230 countries and territories, and the mobile apps in more than 150 countries.

• On the web at gemini.google.com, with any Google account.
• On Android, where Gemini is now the phone's assistant (more on that below), and in the Gemini app on iPhone.
• On your computer, in desktop apps for Mac and Windows. Google's release notes say the Windows app arrived on September 10, 2026, and opens with Alt + Space.
• In Gmail, Docs, Sheets and Meet, as a side panel, on paid personal plans and most Google Workspace plans.
• In Chrome, where it can read the page you're on. A newer auto browse feature completes web tasks for you, for Google AI Pro and Ultra subscribers who are 18 or older and in the US.
• On watches, in cars and at home: on Wear OS watches, in Android Auto and cars with Google built-in, and on Google Home and Nest speakers through an opt-in early-access program.

Does Gemini replace Google Assistant?

On Android phones, yes. Google said in March 2025 that it would move people from Google Assistant to Gemini, starting with phones and following with tablets, cars, headphones and watches. In an email to users reported by 9to5Google, Google said it would start removing Assistant from phones, tablets, Wear OS watches, headphones and Android Auto on September 4, 2026, and that you can't switch back afterward.

Two places were left out of that cutoff: cars with Google built-in, which keep Assistant for now, and Google Home speakers and displays, which move to Gemini only if you opt in. Google says a home that switches can't go back, so if your smart-home routines matter to you, test them on one speaker first.

What can Gemini actually do?

Gemini can answer questions, write and edit text, summarize files, research a topic across many websites, make and edit images and videos, hold a live voice conversation, and read your Gmail and Google Calendar when you connect them. Most of these work on the free plan, with lower limits.

Writing, summaries and files

This is what most people use it for. Paste in a messy draft and ask for a cleaner one, or upload a contract and ask what it commits you to. Google's upload help page allows up to 10 files per prompt. Most files can be up to 100 MB and videos up to 2 GB, with a total video length of five minutes on the free plan and an hour on Pro or Ultra.

Here's a before-and-after. Say Darnell is an operations manager at a 40-person plumbing company in Indianapolis, and he needs to tell the crews about a schedule change.

Write an email about the new schedule.

Write a short email from Darnell to our 22 field technicians. Starting Monday, November 2, the first job starts at 7:30 a.m. instead of 8:00, and trucks must be back by 4:30 p.m. Keep it under 120 words, friendly and direct. End by asking anyone with a conflict to reply by Thursday.

The second prompt gets an email he could send with one read-through. The first gets a generic memo he'd have to rewrite.

Deep Research

Deep Research is Gemini's research mode: it browses up to hundreds of websites, and your Gmail, Drive and Chat if you allow it, then writes a multi-page report with sources. It's on the free plan, with fewer runs.

Say Colleen runs marketing for a regional credit union in Ohio and wants a quick view of how competitors price their checking accounts. A Deep Research run gives her a first draft with links in minutes. She still has to click through and check the numbers, because the report is only as good as the pages it found.

Canvas, images and video

Canvas opens a document or code editor next to the chat, and can turn what you've written into a web page, an infographic, a quiz or a spoken Audio Overview. For images, Google's about page names its Nano Banana image models, which can create pictures and edit photos you upload. Video generation is on the paid plans, using Google's newer video model, Gemini Omni.

Gemini Live

Gemini Live is a back-and-forth voice conversation. You can interrupt it, change direction mid-sentence, and on a phone share your camera or screen so it can see what you're looking at. It's the most natural way to use Gemini while walking or driving.

Connected apps: Gmail, Calendar and more

Connected apps let Gemini read and act in other apps when you ask, such as finding an email in Gmail or checking your Google Calendar. Google's help page lists Gmail, Google Calendar, Google Home, YouTube Music and Spotify, plus Phone, Messages and WhatsApp on Android, and the release notes say more than 40 apps were available as of October 13, 2026. To turn one on:

1. Go to gemini.google.com.
2. In the sidebar, click Customize (click More first if you don't see it).
3. Click Connectors.
4. Find the app and click Add.
5. In a chat, type @ and pick the app to use it.

@Gmail find the last three emails from our landlord about the lease renewal and list every date and dollar amount they mention.

Skills, agents and the Daily Brief

Gemini is starting to do things on its own, not just answer. Google's release notes say saved instructions called Gems are being turned into "skills" that Gemini can also build from your chats. Google also announced a Daily Brief, a morning digest from your Gmail and Calendar for paid subscribers starting in the US, and Gemini Spark, an always-on agent in beta for US Ultra subscribers.

Gemini Notebook (formerly NotebookLM)

NotebookLM, Google's tool for asking questions about your own documents, is now called Gemini Notebook, and Google's help center uses the new name. It's still a separate app, and your notebooks sync with the Gemini app. If you study, research or prepare long reports from a set of sources, it's worth a look.

How much does Gemini cost?

Gemini is free to use, and paid Google AI plans cost $4.99, $19.99, or $99.99 to $199.99 a month in the US. Each paid plan also includes Google storage shared across Gmail, Drive and Photos. These are the prices on Google's subscription page in October 2026:

Google says limits refresh every five hours up to a weekly cap, and depend on how heavy your prompts are. The paid Google AI plans are for people 18 and older.

Which plan should you get?

Start free. Most people can tell within a week whether they hit the limits. If you do, my pick for most people is Google AI Pro: at $19.99 it's priced like other leading assistants and includes 5 TB of storage, which many people would pay Google for anyway. Google AI Plus makes sense if you mainly want more storage and Gemini inside Gmail. Ultra is for people who use it all day and want the agent features first.

Gemini for work accounts

If your company uses Google Workspace, Gemini may already be included. Google's Workspace pricing page lists Gemini in Gmail and the Gemini app on Business Starter, and Gemini in Gmail, Docs, Sheets, Meet, Chat and Vids on Business Standard and above. Ask your admin before you buy a personal plan just to use it at work.

Is Gemini private? The settings to check

Gemini isn't private by default on a personal account. With the Keep Activity setting on, Google's privacy hub says your chats can be used to improve Google's services, including training its AI models, and human reviewers may read a sample of them.

The details worth knowing, all from that page:

• Activity is deleted after 18 months by default. You can change that to 3 or 36 months, or turn auto-delete off.
• Turning Keep Activity off stops future chats from being used for training unless you send feedback, but Google still keeps them for 72 hours.
• Chats picked for human review are disconnected from your account first and can be kept for up to three years, even if you delete your activity.
• Temporary chats are kept for 72 hours and aren't used to train Google's models.

Work accounts are different. Google's Workspace privacy hub says your organization's content isn't used to train models outside your organization without permission.

My rule is simple: on a personal account, don't paste anything into Gemini you wouldn't paste into an email to a vendor. Say Kevin is an HR coordinator at a 120-person logistics firm in Memphis drafting a termination letter. The structure and tone are fine to work on in Gemini. The employee's name, salary and medical leave details belong only in the company's approved tools.

What Gemini gets wrong

Gemini can be wrong, and Google says so: its privacy hub warns the apps may produce inaccurate or offensive information, and its about page asks you to check responses for accuracy. Treat it like a fast, well-read assistant on its first week, not an expert.

• Check numbers, dates and names against the source before they go into anything you send.
• Ask for sources in research and open them. A link that doesn't say what the summary claims is a red flag.
• Expect features to vary. Many depend on your country, language, device, age and plan, so a feature in this article may not be on your account yet.

Gemini for developers: the API in one paragraph

Developers can build Gemini into their own apps through the Gemini API, with a key from Google AI Studio. Google's API pricing page shows a free tier for many of the smaller models, where your data may be used to improve Google's products, and pay-per-use pricing for the rest, including Gemini 4 Argon. You don't need a Google AI subscription to use the API, and the subscription doesn't include API usage.

How to get started with Gemini

Getting started takes about five minutes. Here's the order I'd do it in:

1. Go to gemini.google.com or install the app, and sign in with your Google account. Google's age rules require you to be 13 or older on a personal account (or the minimum age in your country) and 18 or older on a work account.
2. Open the privacy settings and decide on Keep Activity before you type anything personal.
3. Connect Gmail and Google Calendar under Customize, Connectors if you want it to work with your own information.
4. Try three real tasks from this week: an email, a question you'd normally search, and a document you need summarized.
5. Use it free for a week. Upgrade only if you keep hitting the limits.

If you're also weighing ChatGPT, my Gemini vs. ChatGPT comparison goes through prices, features and which one I'd pick for different kinds of work.