For two weeks, Google's most powerful AI model was invitation-only. When Google unveiled Gemini 4 Argon on September 30, only a small group of approved cybersecurity testers could use it. On October 13, Google opened it up to the world.
Developers can now build with it through the paid tier of the Gemini API. Businesses can buy it through Google's enterprise products. And if you pay for Google AI Pro or Ultra, you'll find it in the Gemini app and in AI Mode in Search.
Built for the long jobs
Argon isn't designed for quick questions. It's built for work that takes many steps: digging through huge codebases, running research projects, or analysing a business from top to bottom. It can hold about a million tokens at once, which is roughly several novels' worth of text, so it can keep a whole project in mind while it works.
If you're a developer, the clock is ticking
Launch pricing is $2 per million input tokens and $10 per million output tokens, but only until November 30. From December 1, prices double. If you're curious, this is the cheapest it'll be to try.
The bigger shift
The AI race is changing. A year ago, the question was which company had the smartest chatbot. Now it's whose AI can actually finish a long, real-world job on its own. Argon is Google's answer, and the fact that it went to security testers first shows how carefully the biggest labs are now releasing their most capable models.
Get stories like this in your inbox every week. Subscribe to The Week in Perspective.
Source: Google