Google Releases Gemini 4 Argon: What It Can Do and Who Can Use It

On October 13, 2026, Google unveiled Gemini 4 Argon, a new advanced model in the Gemini family for complex tasks. In the Gemini app, it is available to Google AI Pro and Ultra subscribers, and developers and companies can connect to it through the API.

Argon is built for complex, multi-step tasks

According to Google, the model is built for programming, finance, legal work, cybersecurity and science. The company says Argon ranks first on the Vals Index, which measures how models perform in finance, programming, law and tax tasks. These figures come from Google itself.

The Gemini app gets a tasks mode and connected services

• Answers are more direct, with fewer caveats and less repetition.
• During long tasks, the model more often reports on its progress.
• A “chat — tasks” switch: in tasks mode, Gemini keeps working in Google services even when you are away from your computer.
• Connections to more than 40 apps, including Instacart, OpenTable, Adobe and Shopify.
• The Gemini app for macOS includes a built-in programming interface, Google Antigravity.

Developer pricing is introductory for now

Through the API, Argon launched at an introductory price: $2 per million input tokens and $10 per million output tokens, with cached input tokens 95% cheaper. Starting November 30, 2026, the price will be $4 and $20, respectively. The model’s limit is 1 million tokens per request and response combined.

Argon is only the beginning: Google promises to release the rest of the Gemini 4 family later.

Frequently asked questions