# This week in AI — Gemini Tasks goes free, Gemini 4 Neon, GPT-6.2, and AI that pays for you

> Official announcements only, Nov 13–19, 2026. Gemini's free Task mode in six countries, Google's Gemini 4 Neon and its API pricing, OpenAI's GPT-6.2 Astra and Sol, ChatGPT Wallet, and Genie 4.

- URL: https://aiweekly.app/en/blog/ai-news-2026-11-w3/
- Published: 2026-11-20
- Publisher: AI Weekly (https://aiweekly.app)

## Key takeaways

From official announcements, Nov 13–19, 2026: Google rolled out Task mode in the Gemini app for free in the U.S., Japan, Korea, Canada, Australia and India, and released Gemini 4 Neon for AI Pro and Ultra subscribers and the API at $3 per 1M input tokens and $15 per 1M output tokens. OpenAI introduced GPT-6.2 Astra and Sol and ChatGPT Wallet, which lets agents complete purchases with user approval. Google DeepMind opened Genie 4 to eligible AI Ultra subscribers.


We only cover what companies published on their **own blogs and newsrooms**, and every item links to the original. Benchmark claims below are the companies' own numbers, not independent tests.

OpenAI's pages blocked our automated reader this week, so for OpenAI items we stick to **the summary in OpenAI's official RSS feed** and go no further.

## 1. Gemini's Task mode is free in six countries (Nov 17)

Google made it free to **hand Gemini a job, not just a question**. Flip the toggle on the Gemini app's home screen from Chat to **Task**. It runs on Gemini 4 Hydrogen.

- **Where:** rolling out free in the **U.S., Japan, Korea, Canada, Australia and India**. Advanced features start in the U.S.
- **What it does:** Google's examples are scheduling appointments, analyzing utility bills, shopping and ordering groceries. It connects to more than 60 popular apps.
- **More ways in (U.S. first):** text Gemini on Android or iOS (needs a U.S. number on an eligible carrier), or forward an email to gemini@google.com.
- **Payments:** built on the Agent Payments Protocol (AP2) with Mastercard and Visa; American Express is "coming soon." You can **set a spending cap**.
- **Guardrails:** each task runs in a hardware-isolated virtual machine dedicated to your account. Gemini **asks before it clicks "Buy,"** and some sites hand the final click back to you. You can disconnect apps and delete chat history. Calling businesses isn't available in Indiana, Louisiana, Minnesota, Montana or Nebraska.

**What it means for you:** This is the one item this week most people can try today, for free. Set the spending cap before you let it near your card.

Source: [Put Gemini to work for you — Google](https://blog.google/innovation-and-ai/products/gemini-app/tasks-in-gemini-free-new-features/)

## 2. Google launches Gemini 4 Neon (Nov 18)

Google released **Gemini 4 Neon**, joining Carbon (long, hard reasoning) and Hydrogen (fast everyday work). Google leads with coding, finance, legal, tax and security results.

- **Who gets it:** Google AI **Pro and Ultra** subscribers, from the model picker in the Gemini app. Developers on the paid tier of the Gemini API in Google AI Studio.
- **API pricing:** **$3 per 1M input tokens**, **$15 per 1M output tokens**, and $0.15 per 1M cached input tokens (a 95% discount). 1M-token context window.
- **The claims:** 32% lower cost per task than Carbon at the default thinking level, #1 on the Artificial Analysis Coding Agent Index (58.6), and #1 on Legal Research Bench and Tax Agent Bench — all reported by Google.
- A cyber-focused version is going only to vetted defenders in Google's Fairwind Program.

**What it means for you:** If you pay for Gemini, try it from the model menu. If you build on the API, heavy prompt caching is where the pricing gets interesting.

Source: [Introducing Gemini 4 Neon — Google](https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-neon/)

## 3. OpenAI introduces GPT-6.2 Astra and Sol (Nov 17)

Per OpenAI's RSS summary, **GPT-6.2 Astra** and **GPT-6.2 Sol** bring better professional outputs, coding and computer use, with **more efficiency at lower cost**, across ChatGPT Work, Codex and the API.

**What it means for you:** The summary doesn't say whether free ChatGPT users get them or what the API costs. If you use ChatGPT Work or Codex, check your model list.

Source: [Introducing GPT-6.2 Astra and Sol — OpenAI](https://openai.com/index/introducing-gpt-6-2-astra-sol)

## 4. ChatGPT Wallet lets agents check out for you (Nov 13)

Per OpenAI's RSS summary, **ChatGPT Wallet** lets agents in ChatGPT Work, Codex and dots **complete purchases with your approval**, keeping your existing card protections in place.

**What it means for you:** Google and OpenAI both shipped "AI that pays" within a week. Whichever you use, keep the approval step on. The summary doesn't mention countries or pricing.

Source: [Introducing ChatGPT Wallet — OpenAI](https://openai.com/index/introducing-chatgpt-wallet)

## 5. Genie 4: worlds you can step into (Nov 18)

Google DeepMind opened **Genie 4**, its world model, through a new browser experience called **Project Genie**.

- **Who gets it:** eligible **Google AI Ultra** subscribers. Pro members can join a waitlist on the same page.
- You can prompt mid-session to change the weather, add characters or redirect the action, and it generates **native audio in real time**. There's also an agent that builds worlds from plain language, and a community gallery for publishing and remixing.
- Genie 3 sessions now run up to **3 minutes** per generation, up from 60 seconds.
- All video and audio carries **SynthID watermarking**. Google lists camera latency, physical interactions, audio precision and scene permanence as open problems.

**What it means for you:** Ultra-only for now, so most people will be watching from the waitlist. Worth knowing: anything it makes is watermarked as AI-generated.

Source: [Genie 4 is now available — Google](https://blog.google/innovation-and-ai/models-and-research/google-deepmind/genie-4-now-available-project-genie/)

## Also this week

- **Gemini on low-memory Android phones (Nov 18):** Google re-architected the Gemini app to run on eligible Android (Go edition) devices, with Gemini Live in 45+ languages, camera-based translation and image creation. [Source — Google](https://blog.google/innovation-and-ai/products/gemini-app/gemini-lower-memory-devices/)
- **Child safety tools (Nov 16):** Per OpenAI's RSS summary, new tools, guidance and partnerships to prevent, detect and respond to online child sexual exploitation and abuse. [Source — OpenAI](https://openai.com/index/strengthening-online-child-safety)

## The takeaway

AI is moving from answering to acting — and this week, to paying. The more you delegate, the more it matters to know what to hand off and what to check yourself. [Find out your AI type](https://aiweekly.app/en/level/) in 3 minutes.

## Sources

- [Put Gemini to work for you — Google (2026-11-17)](https://blog.google/innovation-and-ai/products/gemini-app/tasks-in-gemini-free-new-features/)
- [Introducing Gemini 4 Neon — Google (2026-11-18)](https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-neon/)
- [Introducing GPT-6.2 Astra and Sol — OpenAI (2026-11-17)](https://openai.com/index/introducing-gpt-6-2-astra-sol)
- [Introducing ChatGPT Wallet — OpenAI (2026-11-13)](https://openai.com/index/introducing-chatgpt-wallet)
- [Genie 4 is now available — Google (2026-11-18)](https://blog.google/innovation-and-ai/models-and-research/google-deepmind/genie-4-now-available-project-genie/)
- [The Gemini app is now available on lower-memory Android devices — Google (2026-11-18)](https://blog.google/innovation-and-ai/products/gemini-app/gemini-lower-memory-devices/)
- [Strengthening the fight against online child sexual abuse with new tools and support — OpenAI (2026-11-16)](https://openai.com/index/strengthening-online-child-safety)
