What is Gemini? Everything you should know about Google’s AI tool

since then ChatGPT This technology first made waves in late 2022, and tech companies have been racing to make their mark in the field of generative AI. Google, which has long been a powerhouse in artificial intelligence, has focused its efforts on… twin system.

Gemini was launched in December 2023 by Google DeepMind, and has since gone through several model generations, such as Gemini 3.5 In May 2026 and Latest model Flash 3.6released in July.

So what is Gemini? What started as Great language model It evolved into something much broader. The name now refers to Google’s family of AI models, as well as its consumer AI assistant.

These models can understand text, images, audio, video, and code and answer questions about them.

Here’s how Gemini works, what he can do and where his limits still appear.

What can Gemini do?

Gemini has evolved beyond just a standard question and answer chatbot Since its launch.

The current model lineup includes Gemini 3.6 Flash for faster everyday tasks, programming and multi-step work, and 3.5 Flash-Lite for the quickest answers. Gemini 3.1 Pro handles more complex thinking and analysis.

It can connect to Google services such as gmail, calendar, Docs, Sheets, and Drive. Depending on the permissions you grant, you can ask it to find information, Summarize files Assist in coordinating work across applications.

Gemini also includes many built-in features for searching, writing, programming, and customization.

Google/Screenshot by CNET

Deep Research can search multiple sources, follow a research plan, and turn what it finds into a report with links you can check. Canvas is designed for editing alongside an interactive panel within Gemini for drafting long documents or optimizing code.

Precious stones It allows you to create custom versions of Gemini with their own instructions and focus.

Google/Screenshot by CNET

Gemini Live provides instant hands-free conversations with 10 new voice options, even when the app is in the background or your phone is locked. You can also pause and resume conversations whenever you want, which is a great feature.

You can too Share your phone camera Or screen with Gemini Live and ask questions about what you see. For example, you could show them a device and ask them what the button does, or share your screen while comparing products online.

Google also adds More proxy featuresWhich means that Gemini can carry out longer sequences of tasks rather than just telling you how to do them. Gemini Spark It can handle scheduled tasks and multi-step workflows via connected Google apps, though access remains limited to certain users and plans.

Where you can use Gemini

So, how does Gemini play a role in the devices you use every day? For starters, it’s integrated directly Google Pixel phones and serves as the primary Google AI Assistant across many Android devices. It can help you write messages, summarize information, answer questions about what’s on your screen, and work with connected Google apps.

You can also use Gemini through its web, Android, iOS and Mac Apps, as well as through supported Chrome and Google Workspace accounts.

Gemini powers Google Home features that can summarize what the camera or doorbell recorded, search past footage, and answer questions about the live stream. It’s another trade-off between comfort and… Exposure to privacy.

Then there Overviews of artificial intelligence In Google search results. If you notice more detailed answers appearing at the top of your searches, it’s because of this integration. Gemini helps analyze complex topics by combining information from several sources into bite-sized explanations.

Google has come under fire for some strange AI Overviews tips for users during its US rollout in 2024, including suggesting people Eat rocks daily Or put glue in pizza recipes. Company quickly Made some repairs.

AI Overviews is now available in more than 200 countries and regions, in more than 40 languages, and in more than 100 languages 2.5 billion monthly active users (Although AI Overviews appears by default, active users may just be people searching for information on Google as they always do.)

Artificial intelligence fashionan in-search conversational experience designed for longer questions and follow-ups, has surpassed 1 billion monthly active users. Gemini 3.5 Flash now plays it by default worldwide.

Some people still don’t like the AI ​​overviewThe downside is that you can’t disable it. However, we’ve got you covered a Some solutions that may help however.

Create photos, videos and music with Gemini

When Gemini was first launched, it didn’t take long for things to go sideways. Google has faced some criticism Hallucinations How were historical figures and different races depicted? She made headlines for posting photos of black and Asian Nazi soldiers. Critics accused Google of trying too hard to show diversity. Google has shut down its Gemini image generation process, promising to fix the mess.

Uses existing image models Nano banana name and can make targeted changes such as replacing objects, adjusting backgrounds, and combining elements from multiple images.

Google/Screenshot by CNET

Gemini can also create videos and music. Uses Google Show and Gemini Omni To create and edit video, while Liria Models create music from text or picture prompts. Access and creation limits depend on subscription plan and location.

Privacy and your Gemini conversations

Since Gemini can work with connected Google profiles and apps, you should check your privacy settings before using it for potentially sensitive information.

For personal accounts, Google may use Gemini activity to improve its services, including with the help of human reviewers, when Keep active The setting is turned on. Google advises you not to enter confidential information that you do not want to be reviewed or used to improve its products.

You can turn off Keep active Or start a tentative conversation. Ephemeral chats do not appear in your recent conversations and are not used to improve Google’s AI with the help of human reviewers. Just keep in mind that Google still retains chats for up to 72 hours to provide service, respond to issues, and protect users.

Google/Screenshot by CNET

Connected apps have separate permissions, which you can review or revoke as well.

Pricing and access

Gemini is free as your personal AI assistant. The free version currently includes access to Gemini 3.6 Flash, 3.5 Flash-Lite, limited access to 3.1 Pro, image generation and editing, Deep Research, Gemini Live, Canvas, and Gems. Usage limits can change based on demand, feature, and location.

For higher limits and additional features, Google currently offers three main features Subscription plans:

• Google AI Plus costs $5 per month.
• Google AI Pro costs $20 per month.
• Google AI Ultra starts at $100 per month.

Pricing, introductory offers, storage, and feature availability vary by country.

For developers and businesses, Google has set up a tiered pricing structure for its Gemini API models, such as Flash, Omni, and Pro. Developers can access it through Google AI Studio or use Vertex AI through Google Cloud to integrate Gemini’s capabilities into their applications.

Both models offer scalable use of AI, with price caps and rates varying by tier and token length. You can check detailed pricing rates at Official website of Google. There’s also a free tier, giving you a taste with limited usage, which is great for testing the waters before diving in.

Read on to check CNET’s full review of Gemini.