Gemini 4 Argon is Google's newest and most capable Gemini model, launched on 13 October for complex, multi-step work. To use it, subscribe to Google AI Pro or Ultra, open the Gemini app or AI Mode in Search, and give it a meaty task: a report to analyse, a deck to build from your files, or a job to run in tasks mode while you're away. Developers reach it as gemini-4-argon in AI Studio and the Gemini API.
Google released Gemini 4 Argon on 13 October and pitched it squarely at harder work: long reports, multi-step projects and jobs that used to need a person sat at the keyboard. If you run a small business, the useful question isn't how it scores on a benchmark. It's what you can hand to it on Monday morning, and how to set it up so the answers sound like you.
This is a practical walkthrough: who can use Gemini 4 Argon, where to find it, the new tasks mode, and prompts built for small business jobs. The facts all come from Google's own announcement, the Gemini app release notes and the Gemini API docs.
Who can use Gemini 4 Argon, and where
| Where | Who gets it |
|---|---|
| Gemini app | Google AI Pro and Ultra subscribers, rolling out worldwide |
| AI Mode in Google Search | Google AI Pro and Ultra subscribers |
| Google Antigravity | Google AI Ultra subscribers |
| AI Studio and the Gemini API | Developers, as the model gemini-4-argon |
| Gemini for business | Enterprises |
So the short version is that you'll need a Google AI Pro or Ultra subscription to use it in the Gemini app. If you're on the standard free Gemini, you won't see it there.
What's different about Argon
Google describes Argon as its latest frontier model, built for complex, multi-step work. Day to day, three changes stand out.
- Straighter answers. Google says it answers more directly, with less hedging, less repetition and fewer cliches. That's a welcome change if you've ever had to cut three paragraphs of preamble out of an AI draft.
- Progress updates on long jobs. On longer tasks it shares where it's up to more often, so you can see it's on track before it finishes.
- Better at following style instructions. Google says it sticks to style instructions better than earlier Gemini models, which matters when you want copy in your own voice.
Google's release notes give three example jobs: analysing a complex financial report, building a presentation from multiple sources, and creating an interactive 3D model. In the announcement, Argon also built a client proposal deck from a folder of project files.
How to start using Gemini 4 Argon in the Gemini app
- Make sure your Google account has a Google AI Pro or Ultra subscription.
- Open the Gemini app on the web, on your phone or on the desktop app, and start a new chat.
- Check the model you're using in the chat before you begin. Argon is rolling out, so it may reach your account a little after launch day.
- Give it the whole job in one go: what you want, who it's for, the files or sources to use, and what a good result looks like.
- For longer jobs, switch on tasks mode (below) so it can keep going when you step away.
Switch from chat to tasks
This is the change I'd try first. Google added a toggle in the main chat window that switches Gemini from chat to tasks. In tasks mode, Gemini can keep working when you're away from your computer, across your Google apps.
Pair that with connected apps. Gemini can now connect to more than 40 apps and services and finish tasks inside one conversation. Google names Instacart, OpenTable, Adobe and Shopify among them.
Set it up to sound like you
Open the side menu and choose Customize. From one place you can shape how Gemini responds, connect your everyday tools and add skills for repetitive jobs, such as writing in your style or following your design rules. If you haven't built any yet, my guide to creating Gemini skills takes you through it, and Gemini Gems are still handy for saved, reusable helpers.
Prompts to try with Gemini 4 Argon
Argon earns its keep on bigger, messier jobs. Give it context, files and a clear finish line. These prompts are written for small business owners; swap in your own details.
Copy this prompt
Read the attached report: [upload your sales, website or finance report].
Tell me the three things that changed most, why they probably changed, and what I should do about each one over the next few weeks.
Write it for someone who runs the business, not an analyst, and keep it to one page.
Copy this prompt
Use the files in [folder or attachments] to build a client proposal deck for [client name].
Their goal: [what they want]. My offer: [what you'll do].
Keep it to ten slides, lead with their problem, and match the tone of my past proposals.
Copy this prompt
Here are three examples of my writing: [paste them].
Study the tone, sentence length and words I use and avoid.
Now write a short welcome email for new customers of [business] in that same style, with one clear next step.Work with meWant AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Copy this prompt
Plan a full week of content for [business] from this one idea: [idea].
Give me a blog outline, three social posts, a short email and a video script.
Tell me which piece to publish first and why.
A tasks mode prompt for a job that runs while you're away
Copy this prompt
Go through my inbox for messages from customers in the last week.
Group them by topic, draft a reply for each one that needs an answer, and save the drafts without sending.
When you're done, give me a short summary of what needs my decision.
Getting better results from Argon
- Give it the files, not a summary. Argon handles reports, PDFs and folders of documents, so hand over the source material and let it do the reading.
- Describe the finished result. Say how long it should be, who it's for and what the reader should do next. A clear finish line gets a tighter answer.
- Use tasks mode for anything with several steps. Let it work through the job and check its progress updates rather than babysitting each step.
- Keep your voice in Customize. Style notes saved once apply every time, so you're not pasting the same instructions into each chat.
For developers: Gemini 4 Argon in AI Studio and the API
If you build your own tools, the model code is gemini-4-argon. Google's docs list text, image, video, audio and PDF as inputs and text as the output. It supports function calling, search grounding, code execution, structured outputs, URL context and thinking levels of low, medium and high. It doesn't generate images or audio and doesn't support the Live API.
Tools that work with Gemini 4 Argon
Argon gets far more useful once it can reach the places your work already lives. These are the ones worth connecting or pairing first:
- Gmail, Google Drive, Google Docs and Google Calendar, since tasks mode works across your Google apps.
- Shopify, one of the connected apps Google names, for store owners.
- Adobe, also named by Google, for design and creative jobs.
- OpenTable and Instacart, the other connected apps in Google's own examples.
- Google Sheets for keeping any analysis Argon gives you so you can track it over time.
- Canva for polishing a deck or social plan into finished designs once Argon has drafted it.
New to AI tools altogether? Start with my plain-English guide on how to use AI in your business, then come back to Argon for the heavier jobs.
If you'd like help setting Gemini up for your team, you can work with me here.
Frequently asked questions
Is Gemini 4 Argon free?
Not in the Gemini app. Google says it's available to Google AI Pro and Ultra subscribers in the Gemini app and AI Mode in Search. Developers can reach it through AI Studio and the Gemini API.
When was Gemini 4 Argon released?
Google announced it on 13 October, and it started rolling out worldwide to Google AI Pro and Ultra subscribers in the Gemini app the same day.
What is tasks mode in Gemini?
It's a toggle in the main chat window that switches Gemini from chat to tasks. In tasks mode, Gemini can keep working across your Google apps while you're away from your computer.
Can Gemini 4 Argon make images?
The Gemini API docs list text as its only output and say it doesn't support image or audio generation. It can read images, video, audio and PDFs you give it.
What's Gemini 4 Argon best for?
Complex, multi-step work. Google's own examples include analysing a complex financial report, building a presentation from multiple sources and building a client proposal deck from a folder of project files.
