Google Launches Gemini 4 Argon, Gives Cyber Defenders First Access

Google has launched Gemini 4 Argon, granting early access to trusted cyber defenders to test its security capabilities. The model is designed to assist in complex software engineering, enterprise knowledge work, and cybersecurity tasks. Google introduced Argon on Sept. 30 and is providing early access through its Fairwind Program before a wider release.

This allows trusted defenders to explore the model's full security features without the usual cyber guardrails. Google trained Argon specifically for defensive cybersecurity, enabling it to autonomously find, validate, and patch critical software vulnerabilities. Eventually, broader access will be granted to paid API customers and Google AI Ultra subscribers, with introductory API pricing set at $2 per million input tokens and $10 per million output tokens.

One notable test involved Wiz’s Scan for Good initiative, which uncovered a previously unknown critical vulnerability in healthcare software used by hospitals worldwide. While Argon shows promise in vulnerability research, its performance varies on coding benchmarks. Google is also enhancing safeguards to address potential risks associated with misuse and prompt injection, and is collaborating with the US government for pre-release safety evaluations.

The phased rollout is intended to bolster confidence in the model's rollout while making defensive capabilities available sooner. However, the success of Argon will depend on its reliability in real-world scenarios and the level of oversight and safeguards implemented as access expands.