Google Opens Gemini Tasks Free to Millions in Six Countries, Powered by Gemini 4 Hydrogen

Google is making Gemini Tasks free for hundreds of millions of people in the United States, Japan, South Korea, Canada, Australia and India. Powered by the new Gemini 4 Hydrogen model, the agentic mode turns Gemini from a chatbot into a helper that can call businesses, book appointments, shop and pay through Google Pay, and sign into websites on your behalf.

Nov 19, 2026 - 04:15
 0  3
Google Opens Gemini Tasks Free to Millions in Six Countries, Powered by Gemini 4 Hydrogen
A hand holds a Samsung Galaxy smartphone displaying the Google Gemini app welcome screen, which reads "Gemini is your new AI assistant from Google", at a product display table with smartwatches in the background.

Google Opens Gemini Tasks Free to Millions in Six Countries, Powered by Gemini 4 Hydrogen

Google is turning Gemini into a far more hands-on helper. The company announced on November 17 that Tasks, the agentic mode inside the Gemini app, is now free for hundreds of millions of people in the United States, Japan, South Korea, Canada, Australia and India, running on its new Gemini 4 Hydrogen model. Instead of only answering questions, Gemini can now carry out the errand itself: it can ring a business, book the appointment, compare prices, and pay for the order.

Josh Woodward, the vice president who leads the Gemini app, put it plainly in Google's announcement: "Gemini can now make the calls, pay the bills, and finish your tasks." It is the biggest step yet in Google's push to make Gemini an agent that acts, not just a chatbot that chats.

From a test feature to a free assistant

Tasks did not arrive out of nowhere. Google first tested the idea in May 2026 with a small group of users under the name Gemini Spark, a round-the-clock agentic helper. It was later opened to subscribers of the company's Google AI Ultra plan in the United States, and the feature eventually became known as Task Mode inside the app.

The November 17 change removes the paywall. Anyone using Gemini 4 Hydrogen in the six launch countries can switch it on, and the advanced capabilities that make the launch newsworthy, the phone calls and the checkout, are rolling out first in the United States. Getting started is deliberately simple: open Gemini and flip the home-screen toggle from Chat to Tasks, then describe what needs doing. In the US, users can also text Gemini or forward emails to gemini@google.com to set a task in motion.

What Gemini Tasks can actually do

Calling businesses and booking appointments

The headline capability is the phone. Gemini can find nearby businesses through Google Maps, call several of them at the same time, and book an appointment on the user's behalf, adding it straight to Google Calendar. Woodward's own example was a family errand: with the check-engine light on in the family's 2019 Honda Odyssey minivan, he asked Gemini to call the two nearest dealerships and find an early Friday morning service slot. Gemini sat through the hold music, found a time he was free, and had the appointment booked in under five minutes.

On the other end of the line, Gemini says straight away that it is an AI, and businesses can opt out of receiving such calls whenever they want. Google also says Gemini will not hand over payment or private account details unless the user explicitly allows it. The calling feature is limited to local businesses in the United States for now, and it is not available in Indiana, Louisiana, Minnesota, Montana or Nebraska.

Shopping and paying through Google Pay

Shopping is the second major piece. Gemini can research products using Google's Shopping Graph, compare options, hunt for discount codes, and complete the purchase through Google Pay. In one example from the announcement, Gemini compared USB-C chargers, asked follow-up questions about wattage and colour, found a discount code, and saved Woodward an extra 10 dollars at checkout.

Payments run on Google's Agent Payments Protocol, known as AP2, and are protected with digital tokens on supported Mastercard and Visa cards, with American Express support coming later. Cardholders still earn their usual points, cashback and purchase protections, and receipts and tracking details land in Google Wallet, which is expanding to iOS from the same day. Users can also set a hard spending cap in advance, and Gemini is designed to stay inside that budget. Before it clicks buy, Gemini is designed to ask for confirmation, or for advance approval, so a purchase never goes through silently. Google says the Universal Commerce Protocol will be integrated into Gemini next, to make the whole journey from discovery to checkout faster and more reliable.

Signing in and sorting out bills

Tasks can also reach into the places where a user's life already lives. With the new Google Password Manager integration, Gemini can sign into websites securely when asked, and analyse documents such as utility bills. Woodward described a surprise 389-dollar electric bill: after getting permission to access the utility portal and past statements in his email, Gemini traced the problem to the family's heat pump switching on around 4 p.m., right in the middle of peak pricing, something he said they would never have spotted digging through PDFs themselves.

The same reach extends to everyday chores. Gemini now connects with more than 60 popular apps, and developers can apply to join the Connected Apps Platform. In the announcement, Woodward texted Gemini to add lasagna, strawberries, cucumbers and vanilla ice cream to an Instacart order, then sent a photo of an empty Cheerios box from the pantry as a follow-up, and the groceries were on the doorstep when the family got home. In another example, forwarded school emails were turned into calendar dates, with a carpool clash during basketball practice flagged and a text to another parent offered as a fix.

How Google says it keeps the agent in check

Handing an AI the keys to phone calls, payments and logins raises obvious questions, and Google has built the feature around a set of guardrails. Each task runs in a hardware-isolated virtual machine dedicated to that user's account, using Google Cloud's Agent Compute Engine, so the work happens in a sandboxed environment that Google says its security red teams test regularly. Prompt-injection filters scan search results and web pages, alongside Google Safe Browsing, to limit which links Gemini can open.

Gemini is also designed to pause and check before the steps that matter, asking for confirmation before buying or clarifying when an instruction is ambiguous. For routine jobs the user already trusts, such as adding an item to a grocery cart, that check can be set to "Always allow". Users can adjust task settings, disconnect Google and third-party apps, or delete chat history at any time. Google has published a technical paper on the approach and invited researchers to take part in its AI Vulnerability Reward Program.

What it cannot do yet

Google is candid that the system is not flawless. Tasks can still fail or miss the mark, and the company asks users to flag mistakes with a thumbs down. The way personal assistants move around the web is still evolving: on some sites Gemini gets most of the way through a task and then hands the screen back so the user can click the final button. Google says it is working directly with major websites and security providers on standards that let personal assistants work without being treated like the malicious bots websites have blocked by default for decades.

The rollout itself is uneven by design. The free base version of Tasks is available in all six countries, but the calling, payment and text-message features start in the United States, and calls are limited to US local businesses. That split is worth keeping in mind for anyone outside the US: Tasks will work as an agent, but the most eye-catching powers arrive later.

Why this matters

Gemini Tasks lands in the middle of a race to turn chatbots into agents that do things. OpenAI has moved in the same direction with ChatGPT Wallet, which lets its AI agents shop and check out for users, while United Airlines recently moved the other way, barring AI agents from booking flights. Google's answer leans on the advantage it already holds: the data, the calendar, the email, the passwords and the payment cards are mostly already inside Google's ecosystem, so the agent does not have to ask the user to send them somewhere new.

Whether that convenience wins people over will depend on how reliably Tasks completes the job, and how comfortable users are letting an AI speak, spend and sign in on their behalf. For now, Google has made the bet free for hundreds of millions of people, and the errands are the test.

Sources

  • Google, "How to use Tasks in Gemini for personal assistance", November 17, 2026 — https://blog.google/innovation-and-ai/products/gemini-app/tasks-in-gemini-free-new-features/
  • TechCrunch, "Google expands access to Gemini Tasks, which can help with calls, purchases", November 17, 2026 — https://techcrunch.com/2026/11/17/google-expands-access-to-gemini-tasks-which-can-help-with-calls-purchases/
  • Neowin, "Google makes its always-on AI agent free, adds phone calls, shopping, and more", November 17, 2026 — https://www.neowin.net/news/google-makes-its-always-on-ai-agent-free-adds-phone-calls-shopping-and-more/
  • The Times of India, "Google expands Gemini Tasks to millions of users, adds calling and shopping features", November 18, 2026 — https://timesofindia.indiatimes.com/technology/tech-news/google-expands-gemini-tasks-to-millions-of-users-adds-calling-and-shopping-features/articleshow/135962173.cms

What's Your Reaction?

Like Like 0
Dislike Dislike 0
Love Love 0
Funny Funny 0
Angry Angry 0
Sad Sad 0
Wow Wow 0
Nahid Joy I’m Nahid Hasan Joy, a technology writer, web developer, and digital enthusiast with a strong interest in the ever-changing world of technology.