How to use Gemini: a beginner's guide to the app, models, Deep Research, skills, tasks and plans

Gemini is Google’s AI assistant, and the basics are quick to learn: open it in a browser or the app, type what you need and send it. What trips up new users is the rest: which model you’re getting, why Gems became skills, what the Task button does and which plan you need. This tutorial walks through each one using Google’s own help pages and subscription page as of October 2026.

What you need

• A Google Account. Google’s sign-in page says a personal account needs you to be 13 or older, or the applicable age in your country. Work accounts need a qualifying Workspace edition and an age of 18 or over. Several features in this guide, including Deep Research, skills, tasks and video generation, are 18+ only.
• A browser, phone or computer. Gemini runs at gemini.google.com and in apps for Android, iPhone and iPad, Windows and Mac.
• Nothing to pay, to begin with. The free plan covers most of this guide. We flag the parts that need Google AI Plus, Pro or Ultra.

Step 1: Choose where to use Gemini

Google says some features work at gemini.google.com without signing in, but you need to sign in to save your chats and use everything else, so sign in from the start.

On Android, Gemini is also the phone’s assistant on most devices. If Gemini is your default assistant, Google says you can open it by pressing and holding the power button, touching and holding Home, or swiping up from a bottom corner. To use “Hey Google”, set up Voice Match in Gemini’s hands-free settings.

On a computer, the desktop app opens over whatever you’re doing: press Alt + Space on Windows, or Option + Space on a Mac for a mini chat (Option + Shift + Space for the full window). Google says some features are Mac-only for now. Our guide to the Gemini app for Windows and Mac covers the differences.

Step 2: Start your first chat

According to Google’s help page, on the web:

1. Go to gemini.google.com and sign in.
2. Type your question or request in the text box at the bottom.
3. Optionally, click Add files to attach a file or image.
4. Click Submit.

In the mobile app, type and tap Send, or tap the microphone, speak your prompt and then tap Send. To start over, click New chat at the top left (open the Menu first if you can’t see it).

Click Edit text next to a prompt you’ve sent to change it, then Update, and Gemini rewrites its answer. And for a conversation you don’t want kept in your history, click Temporary chat next to New chat. Temporary chats need a personal account and don’t appear in your recent chats.

Step 3: Pick a model and thinking level

Gemini runs on four models. Google’s help center describes them as:

• Flash-Lite: fast and cost-efficient, for quick answers.
• Flash: a balance of performance and speed.
• Pro: for complex prompts, coding and in-depth problem-solving.
• Gemini 4 Argon: “top-tier intelligence” for complex work. Our Gemini 4 Argon report covers the model Google released on October 13, 2026.

To switch, tap or click the model name at the bottom of the chat box and choose another. Models with a lock icon need a plan upgrade. You can also choose a thinking level: Low (the default), Medium or High, where High takes more care over complex, multi-step prompts. The choice stays with that conversation.

What you can pick depends on your plan:

With auto selection on, Gemini routes each prompt to the model and thinking level it judges most efficient. You can turn it off under Settings, then Usage limits. Bigger models and higher thinking levels use more of your allowance, so beginners can leave auto on.

Step 4: Talk to Gemini with Gemini Live

Gemini Live is a spoken conversation you can interrupt mid-answer, and it works on the free plan. In the Android or iPhone app, tap Live or swipe left at the bottom of the screen, then follow the onscreen setup the first time. In Chrome on a computer, click Ask Gemini at the top of the browser, then Go Live. Our full tutorial on how to use Gemini Live has the steps for each device.

Step 5: Add files and photos

Gemini can read documents, spreadsheets, photos, videos and code. On the web, type your prompt, click Add files and choose Upload Files, or add files from Google Drive, a code folder, a GitHub repository or a notebook. In the mobile app, tap the add icon to take a photo or pick one from your gallery or files.

Google’s limits, from its file-upload help page:

• Up to 10 files per prompt, subject to availability. Each file in an uploaded folder counts.
• Size: up to 2 GB per video and 100 MB for other files.
• Video and audio length: up to 5 minutes of video and 10 minutes of audio in total, rising to 1 hour and 3 hours on Google AI Pro or Ultra.
• Context window: 32K tokens on the free plan, 128K on AI Plus and 1 million on Pro and Ultra. Google says a million tokens is about 1,500 pages of text, so very long documents need Pro or Ultra to be read in full.

On iPhone and iPad, Google notes that the Gemini app can’t yet add Drive files, only local ones.

Step 6: Make images and videos

On the web, click Open sidebar, choose Images, then pick a template or type a prompt. The image model follows your chat model: Flash-Lite uses Nano Banana 2 Lite, and Flash or Pro use Nano Banana 2, which can take several reference images. Subscribers can click More under an image, then Redo with Pro, to regenerate it with Nano Banana Pro. To edit a picture, open Library in the sidebar, click the image (or upload your own), click Chat and describe the change. Editing images is for users 18 and over.

Video needs a Google AI plan on a personal account and is 18+ only. Open the sidebar, choose Create video, type a prompt and, if you like, add one video and up to five images. Google’s help page says videos are made with Gemini Omni, its latest video model, and can take a few minutes.

Step 7: Run Deep Research

Deep Research runs many searches and turns them into a cited report. It’s on every plan, but you must be 18 and signed in, and Google says free users may lose access at busy times. On the web:

1. In the text box, open the Add files menu and choose Deep Research.
2. Optionally, open Sources and add Gmail or Drive. Google Search is on by default.
3. Type your research question and submit.
4. Gemini drafts a research plan. Edit it if you need to, then click Start research.

Reports usually take 5 to 10 minutes, and you can leave the chat while it works. From the finished report you can create an Audio Overview, export it to Google Docs or share it. Pro and Ultra subscribers can use the Pro model for reports and get higher limits, and only Ultra reports include charts and interactive visuals. Our deep research tools comparison sets Gemini against ChatGPT, Claude and Perplexity.

Step 8: Save your favorite prompts as skills

Skills are reusable instructions, such as “always answer in plain English with three bullet points” or a writing-style guide. They’re replacing Gems: Google is removing Gems from personal accounts from November 2026 and says it recreates them as skills automatically, including supported knowledge files. Our news report on Gemini skills replacing Gems has the dates for work and school accounts.

To create one on the web, according to Google’s help page:

1. Click Customize in the sidebar (under More if you don’t see it), then Skills at the top.
2. Pick a template under Recommended, or click Create and choose Create with Gemini or Create manually.
3. Add a name, description and instructions, then click Create.

You can also just describe a skill in a chat and Gemini saves it. To use one, type / in the chat box (or @ if you have the updated shortcut) and pick it, or let Gemini apply it when it fits. Skills need you to be 18 or over, signed in to a personal account, with Keep Activity on.

Step 9: Try notebooks and tasks

Notebooks keep a project’s sources, instructions and chats together. In the Gemini web app, open the sidebar, create a new notebook under Notebooks, name it and add sources. Notebooks sync with Gemini Notebook, the research tool formerly called NotebookLM, and Google says you can add up to 600 sources depending on your plan. Our guide to what NotebookLM is explains the standalone app and its Audio Overviews.

Tasks are for jobs you want Gemini to carry on with in the background. Google’s help page now says it renamed Gemini Spark to Task “to make it easier to understand”. You’ll find a Chat and Task toggle next to the prompt box:

1. Click Task below the text box.
2. Describe the job and when it should run, for example “every Monday morning, summarize my unread newsletters”.
3. Add files with Upload & tools if needed, then click Submit.
4. Check back, since Gemini may ask for input or ask you to take over its browser.

Tasks need Google AI Pro or Ultra, a personal account, an age of 18 or over, Keep Activity on and Personal Intelligence turned on. They work on the web, in the mobile app and in the Mac app, but not in the EEA, Nigeria, Switzerland or the UK. Google calls the feature experimental, so supervise what it does. Our guide to Gemini Spark covers limits and safety.

Step 10: Check your privacy settings

For signed-in adults, Keep Activity is on by default, which saves your chats to your Google Account. Google’s privacy hub says that when it’s on, Google uses your activity to improve its services, including training generative AI models, and that some chats are reviewed by people after being disconnected from your account. Reviewed chats are kept for up to three years, even if you delete your activity.

To change this on the web, open Settings & help, then Activity:

• Click On, then Turn off or Turn off and delete activity to stop saving chats.
• Change auto-delete from the default 18 months to 3 or 36 months, or turn it off.
• Use Delete to remove the last hour, last day, a custom range or all time.

Turning Keep Activity off disables skills, tasks and some notebook features. Even with it off, or in a temporary chat, Google keeps conversations for up to 72 hours. If you want Gemini off your phone or out of Gmail altogether, see how to turn off Gemini.

Which plan do you need?

Limits are compute-based, so long chats, big files, Deep Research and video use more. Google says the limit refreshes on a rolling five-hour basis until you reach a weekly cap, and subscribers who run out can carry on with Flash-Lite. You can check your usage under Settings, then Usage limits. Our Google AI Plus vs Pro vs Ultra guide covers the storage, Flow credits and other perks in each plan.

Tips for getting better answers

• Say what you want back. Give the audience, length and format, for example “a 100-word summary for my manager”.
• Start a new chat when you change topic. Google notes that long chats use more of your limit, because Gemini re-reads the history.
• Attach the source. Upload the PDF or photo rather than describing it.
• Turn repeats into skills. If you paste the same instructions more than twice, save them.
• Check important facts. Gemini can be wrong, so verify numbers, quotes and anything you’ll rely on.

Where to go next

For a verdict on the assistant as a whole, read our Gemini review or visit the Gemini profile. If you’re still choosing between assistants, compare Gemini vs ChatGPT.