On this page
On November 17, 2026 (exact JSON-LD and RSS timestamp: 2026-11-17T17:00:00+00:00, corresponding to November 18, 2026 at 01:00:00 Asia/Shanghai), Josh Woodward, VP of Labs, Gemini app, and AI Studio at Google, published an announcement titled "Put Gemini to work for you," rolling out Tasks in Gemini free of charge across six countries and reaching what Google describes as hundreds of millions of people.
Rather than marking a new foundation model launch, this update expands an agentic application workflow powered by the previously released Gemini 4 Hydrogen model (API model ID: gemini-4-hydrogen). Readers looking for model architecture and token pricing details can refer to our coverage of Google launching Gemini 4 Hydrogen and our Gemini 4 selection and pricing guide.
Six Eligible Countries and the Home-Screen Toggle
Country Availability List: **According to the official blog post, free access to Tasks in Gemini covers **the U.S., Japan, Korea, Canada, Australia, and India. Google has not announced global free availability, and markets outside this six-country list are not included in this rollout.
Switching From Chat to Task: **In the Gemini app, users activate asynchronous multi-step execution by flipping the home-screen toggle from **Chat to Task, allowing Gemini to carry out background actions without requiring continuous back-and-forth prompts.
U.S.-First Advanced Channels and Regional Exclusions
While the core Task toggle is rolling out across all six listed countries, Google specifies clear geographic and carrier prerequisites in the how-to instructions and footnotes:
Email Forwarding and Text Messaging (Starting in the U.S.): **Starting in the U.S., users can forward emails to **gemini@google.com or text Gemini directly after configuring Settings -> Channels -> Set up Text with Gemini and verifying their phone number. Footnote 2 specifies that RCS texting requires a U.S. phone number with an eligible mobile carrier, meaning email and SMS triggers are not simultaneously available across the other five countries.
Calling Exclusions Across Five U.S. States: **Although the post illustrates Gemini placing phone calls to businesses on a user's behalf, Footnote 1 states that calling through Gemini is **not currently available in Indiana, Louisiana, Minnesota, Montana, and Nebraska. In addition, businesses can opt out of receiving AI calls.
Isolated Execution, Payment Controls, and Unconfirmed Details
Agent Compute Engine and Connected Apps: **Tasks run inside hardware-isolated per-account virtual machines (Agent Compute Engine**) and can draw on Personal Intelligence, Password Manager, Shopping Graph, and more than 60 popular apps via the Connected Apps Platform.
AP2 Checkout and Hard Spending Caps: **Commerce workflows rely on the **AP2 checkout protocol in partnership with Mastercard and Visa (with American Express listed as coming soon). Tasks enforce a hard spending cap and require user confirmation or pre-authorization before completing purchases, with digital receipts saved to Google Wallet ("expanding to iOS starting today").
**Fallback Behavior and Undisclosed Specs: **Google notes that tasks may fail or hand the final confirmation button back to the user. The blog page also carries Google's standard footer notice: "This content is generated by Google AI. Generative AI is experimental." Google has not disclosed per-country user counts, free-tier daily task quotas, or whether the Chat-to-Task toggle differs across iOS, Android, and Web clients.
