How to Use Gemini 4 Argon in the Gemini App

Gemini 4 Argon is Google’s frontier model for deep reasoning and long tasks. Google AI Pro and Ultra subscribers can use it in the Gemini app starting October 13, 2026. Free and Plus plans do not include it.

Argon answers more directly, shares progress on longer jobs, and handles multi-step work across connected apps. It supports a 1-million-token input-output limit through a continuation protocol.

Who Can Access Gemini 4 Argon

You need an active Google AI Pro or Ultra subscription on the personal Google account you use with Gemini. Pro is $19.99 per month in the US; Ultra starts higher. Confirm your plan at one.google.com or inside the Gemini app under your profile.

The model also appears in AI Mode in Google Search for the same plans. Developers can call it separately through the paid Gemini API tier with model ID gemini-4-argon. Introductory API pricing is $2 per million input tokens and $10 per million output tokens through November 30, 2026.

Rollout is global but staged. If Argon is missing after you subscribe, update the Gemini app and check again later.

Switch to Gemini 4 Argon

Open the Gemini app on Android, iOS, or the web at gemini.google.com. Sign in with the account that holds the Pro or Ultra plan.

1. Tap or click the model name in the prompt bar at the bottom of the chat.
2. Select Gemini 4 Argon from the list. Paying users see options that include 3.5 Flash-Lite, 3.8 Flash, 3.1 Pro, and 4 Argon.
3. Choose a thinking level if available: Low (default), Medium, or High. Use High for code, long documents, or multi-step analysis.
4. Type your prompt or attach files. Argon accepts text, images, audio, video, and PDFs within the context window.

The picker may sit inside the text box on the web. On mobile it is usually next to the input field. If the model does not appear, the staged rollout has not reached your account yet.

For more on pricing and access across surfaces, see the related guide on Gemini 4 Argon features and pricing.

Use Tasks Mode for Longer Work

A toggle in the main chat window switches from chat to tasks. In tasks mode, Gemini continues working even when you leave the app or computer. It connects to more than 40 apps such as Instacart, OpenTable, Adobe, and Shopify for actions like ordering or booking.

1. Open a new chat or continue an existing one with Argon selected.
2. Flip the toggle from chat to tasks.
3. Describe the outcome you want, including any files or connected-app steps.
4. Review the plan Gemini shares, then confirm so it runs in the background.
5. Check progress later in the chat or through notifications.

Tasks mode is included with Pro and Ultra. It builds on earlier Spark capabilities and keeps you in the loop with more frequent status updates.

Practical Examples

Try these prompts after selecting Argon and setting thinking to High:

• Attach a long PDF and ask: “Summarize the key findings, list open questions, and draft three follow-up emails.”
• Connect Calendar and Docs, then say: “Review my meetings this week, extract action items, and create a shared checklist.”
• Upload a code snippet or project description and request a refactored version with tests.

Argon reduces hedging and repetition compared with earlier models. It is designed for workflows that span research, writing, coding, and connected-app actions.

Tips for Better Results

Keep the Gemini app updated. Staged rollouts mean the model can appear hours or a day after the official launch for some accounts.

Attach the full source material instead of a short summary when the context window allows it. The 1-million-token limit supports entire documents or long recordings.

Review connected-app permissions before running tasks. You stay in control of what Gemini sends or changes.

If you also develop Android apps, Argon is available in Android Studio and Google AI Studio on the paid API tier or Ultra. Separate guides cover those setups.

Conclusion

Gemini 4 Argon brings frontier reasoning into the Gemini app for Pro and Ultra subscribers. Select it in the model picker, raise the thinking level for hard tasks, and switch to tasks mode when you need background progress across Google and third-party apps.

Update the app, confirm your plan, and start with a concrete multi-step request. The model is built for the longer workflows that earlier versions handled less reliably.

Sources: Google Blog announcement (October 13, 2026), Gemini API documentation, Android Developers Blog.