Go to gemini.google.com (or the mobile app), sign in with any Google account, and you can start chatting for free with Gemini 3 Flash. The skills worth learning first: pick the right model from the dropdown, use Deep Research for cited reports, build a Gem for tasks you repeat, and try Canvas to turn a prompt into a document, quiz, or working prototype. The free tier covers casual use; paid plans start at $7.99/month and mainly buy you smarter models and higher limits.
If you searched how to use Google Gemini, you're in good company — tutorial searches for Gemini spiked this month as Google rolled its newest models and agent features into the app. The catch is that Gemini in 2026 is much more than a chatbot: it's a model picker, a research agent, a coding sandbox, an image and video generator, and a voice assistant sharing one interface. This tutorial walks through it feature by feature, in the order a beginner should actually learn them. (Details below were checked in early July 2026 — Gemini changes fast.)
Gemini at a glance
| What it does | Free tier? | |
|---|---|---|
| Chat | Ask anything, upload files and images | ✅ Yes (daily limits) |
| Model picker | Switch between Flash (fast) and Pro (deeper reasoning) | ⚠️ Flash only |
| Deep Research | Autonomous, cited research reports | ✅ ~5 reports/month |
| Gems | Save a custom AI setup you reuse | ✅ Yes |
| Canvas / Create | Turn prompts into docs, quizzes, web pages, code | ✅ Yes |
| Image generation | "Nano Banana" image model | ✅ Basic |
| Video generation | Gemini Omni (text/image → video) | ❌ Paid plans |
| Gemini Live | Natural voice conversation | ✅ Yes |
Step 1: Sign in and learn the model dropdown
Head to gemini.google.com in any browser, or install the Gemini app on iOS or Android, and sign in with a regular Google account. There's nothing to install on desktop and no waitlist.
The first thing to understand is the model dropdown at the top of the chat. In mid-2026 you'll typically see:
- Gemini 3 Flash / 3.5 Flash — the fast models. Great for everyday questions, summaries, and quick drafts. Free users mostly run on Flash.
- Gemini 3.1 Pro — the reasoning model, with a 1-million-token context window (roughly 1,500 pages of text). Choose it for complex analysis, long documents, and multi-step problems. Requires a paid plan for sustained use.
The habit that separates confident users from frustrated ones: match the model to the task. Flash for speed, Pro for depth. If a Pro-level answer comes back shallow, it's often because the prompt ran on Flash.
One more 2026 change worth knowing: Google now meters usage by compute, not a flat prompt count. Long chats, big files, and heavyweight features consume your allowance faster, and limits refresh on a five-hour cycle with a weekly cap. If you hit a wall mid-project, starting a fresh, focused chat genuinely helps.
Step 2: Upload files and ask real questions
Gemini accepts PDFs, images, spreadsheets, and pasted text directly in the prompt box. The practical wins for beginners:
- Summarize a PDF: drop in a contract or report and ask for key points, risks, or a plain-English translation.
- Explain an image: screenshot an error message, a chart, or a form and ask what it means.
- Work with data: upload a CSV and ask for trends, outliers, or a chart.
Be specific about the output you want ("give me a 5-row table comparing X and Y", "explain like I'm new to this"). Vague prompts get vague answers on every AI tool, and Gemini is no exception.
Step 3: Use Deep Research for anything that needs sources
Deep Research is Gemini's standout feature and the one most beginners never find. Pick it from the tools menu, give it a question ("compare mirrorless cameras under $1,000 for travel"), and it browses dozens of sources for several minutes, then returns a structured, cited report. In 2026 you can also upload your own files as sources and convert finished reports into interactive visuals or quizzes in Canvas.
Free accounts get roughly five Deep Research reports a month — enough to learn the feature. Save them for questions where you'd otherwise open twenty browser tabs.
Step 4: Build a Gem for anything you do twice
A Gem is a custom version of Gemini with saved instructions. Click "Explore Gems" → "New Gem", then write instructions the way you'd brief an assistant: what role it plays, what tone to use, what format to answer in.
Good first Gems:
- An email polisher that rewrites drafts in your voice and keeps them under 150 words.
- A study coach that quizzes you from uploaded notes.
- A social caption writer that knows your product and audience. (Writing for platforms with strict limits? Our character limits guide pairs well with this.)
The payoff is consistency: instead of re-explaining your context in every chat, the Gem remembers the brief.
Step 5: Try Canvas and the Create menu
Canvas turns Gemini from a chat window into a workspace. Through the Create menu you can transform a prompt or a pasted document into a formatted doc, an infographic, a quiz, an Audio Overview (a podcast-style discussion of your content), or even a working web prototype — Gemini writes and runs the code, and you refine it conversationally.
For non-programmers this is the closest thing to magic in the app: "make me a simple calculator that converts recipe servings" produces something you can actually click and use within a minute. Developers who want AI in their actual codebase should look at dedicated agents instead — see our Cursor vs Copilot vs Claude Code comparison.
Step 6: Images, video, and voice
Images are generated by Google's Nano Banana model — just describe what you want in the chat. It's strong on editing too: upload a photo and ask it to change the background or combine two images. Free users get basic access; paid tiers raise the quality and volume. (Curious how it stacks up against Midjourney and DALL·E? See our best AI image generators roundup.)
Video arrived for subscribers via Gemini Omni, which turns text or image prompts into short cinematic clips. It rolled out to Google AI Plus, Pro, and Ultra subscribers worldwide in May 2026 — there's no meaningful free video tier yet.
Gemini Live gives you a hands-free voice conversation — tap the waveform icon in the mobile app. You can interrupt it mid-sentence, point your camera at things and ask about them, and use it while driving or cooking. It's free and criminally underused.
What the plans actually cost in 2026
| Plan | Price | What you get |
|---|---|---|
| Free | $0 | Gemini 3 Flash, daily limits, ~5 Deep Research/mo, 15 GB storage |
| Google AI Plus | $7.99/mo | Higher limits, video generation access, more storage |
| Google AI Pro | $19.99/mo | Gemini 3.1 Pro, 1M context, 2 TB storage, Health & Home Premium |
| Google AI Ultra | $99.99–$200/mo | 5x–20x Pro limits, Deep Think, Gemini Spark agent, 20 TB, YouTube Premium |
Advice for beginners: stay on the free plan for your first month. Upgrade to Plus or Pro only when you repeatedly hit limits or clearly need the Pro model's reasoning. Ultra — with its 24/7 Gemini Spark agent and Daily Brief — is aimed at power users and early adopters.
Frequently Asked Questions
Is Google Gemini free to use?
Yes. The free tier at gemini.google.com includes chat with Gemini 3 Flash, file uploads, Gems, Canvas, basic image generation, Gemini Live voice chat, and about five Deep Research reports per month. Paid plans start at $7.99/month and mainly add stronger models, video generation, and higher usage limits.
What's the difference between Gemini Flash and Gemini Pro?
Flash models (3 Flash, 3.5 Flash) are optimized for speed and handle everyday tasks well. Gemini 3.1 Pro is the deep-reasoning model with a 1-million-token context window, better for long documents, complex analysis, and multi-step work. Free users primarily get Flash; Pro requires a subscription for sustained use.
Is Gemini better than ChatGPT?
They're closer than fans of either admit, and the answer depends on what you do. Gemini's edge is Google integration (Gmail, Docs, Drive), Deep Research, and generous free features; ChatGPT's is its plugin/app ecosystem and some workflow polish. We compare them head-to-head in ChatGPT vs Claude vs Gemini.
Can Gemini generate videos?
Yes — Gemini Omni turns text and image prompts into short high-quality video clips, but it's only available on paid plans (Google AI Plus, Pro, and Ultra). For a broader look at the video space, see our best AI video generators guide.
Does Gemini work with Gmail and Google Docs?
Yes. Gemini connects to Google Workspace apps, so you can ask it to summarize email threads, draft replies, find files in Drive, and pull details from Calendar — one of its biggest practical advantages over standalone chatbots.
Conclusion
Google Gemini in 2026 rewards a staged approach: master plain chat and the model dropdown first, then add Deep Research for anything needing sources, a Gem or two for repeated tasks, and Canvas when you want output you can use rather than just read. Everything above except video generation is available on the free plan, so the best way to learn is simply to open gemini.google.com and work through these six steps. And if you're still choosing your primary AI assistant, our ChatGPT vs Claude vs Gemini comparison will help you decide where Gemini fits.