Inside Gemini 4 Argon, the model Google is testing on its own infrastructure first
Google is testing Gemini 4 Argon, a frontier model with a 1 million token output, first with trusted defenders.
Google announced Gemini 4 Argon, a frontier model first offered to trusted cyber defenders through the Fairwind Program rather than the public. Its output limit rises to 1 million tokens from 64,000, with introductory pricing of $2 per million input tokens and $10 per million output tokens. Argon leads reported benchmarks including 77.9% on DeepSWE v1.1, 91.7% on LVBench, and 68% on CWE-bench v1. For approved security use Google relaxes cyber restrictions, while adding prompt-injection hardening and real-time chain-of-thought monitoring that can stop execution.
- Gemini 4 Argon goes first to trusted defenders in Google's Fairwind Program.
- Output limit rises to 1 million tokens from the prior 64,000.
- Introductory pricing is $2 per million input tokens and $10 output.
- Cyber restrictions are relaxed for trusted teams so Argon can find flaws.
- Google monitors chain-of-thought and can halt execution if intent drifts.
- Vendors
- Products
- Fuchsia Zircon
- AI models
- Gemini 4 Argon
From the original article
4 min read at securityaffairs.com
Google unveils Gemini 4 Argon, a frontier AI model built for coding, enterprise work, and autonomous cybersecurity defense, rolling out to trusted testers. Google announced Gemini 4 Argon, and it’s not going straight to the public. It’s rolling out first to a set of trusted cyber defenders through what Google calls the Fairwind Program, which […]
By Pierluigi Paganini · securityaffairs.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.