Google Makes New Gemini 4 Argon Available for Cybersecurity Experts

Just a day after OpenAI announced the release of its new GPT-6.1 model on Tuesday, Google followed up yesterday with Gemini 4 Argon, a new frontier model that’s only available for cybersecurity experts through its Fairwind Program. The news also comes right after the Wall Street Journal reported that OpenAI had scrapped the release of a new GPT-6.1 Astra model due to safety concerns.

“Built to sustain deep reasoning across complex, long-horizon workflows, Argon is fundamentally changing the way we work and build at Google. It delivers frontier performance in complex workflows across real-world software engineering, enterprise knowledge work like legal and finance, and cybersecurity defense. Safely releasing frontier capabilities at this level requires a phased approach,” explained Koray Kavukcuoglu, SVP, Google DeepMind and Chief AI Architect.

According to Google, Gemini 4 Argon beats OpenAI’s GPT-6 Astra and Anthropic’s Claude Fable 5.1 and Opus 5.5 models across various benchmarks such as DeepSWE v1.1 and the Vals Index, which measures AI models’ ability to perform real-world tasks across finance, coding, legal, and other activities. On the security front, Kavukcuoglu claimed that Gemini 4 Argon found “a critical vulnerability exposing sensitive personal information across healthcare software used by hospitals worldwide, identifying a severe risk that previous frontier models had missed.”

While Google will make Gemini 4 Argon available without cyber guardrails for trusted defenders, the company will implement strong protections before making the new frontier broadly available. To prevent the model from going off the rails, which is something that seems to be happening increasingly frequently, the exec also said that Google is “deploying misalignment mitigations that monitor Argon’s chain-of-thought and actions and stop execution when necessary.”