Google said on October 13 that it is beginning a broad rollout of Gemini 4 Argon and expanding its Fairwind Program to more trusted cyber defenders.

The company said the model’s stronger cyber and biology-related capabilities led it to add new safeguards.

In a safety update, Google said Argon solved all 12 end-to-end simulated cyberattack scenarios, up from five in the previous generation.

During limited access, cloud security firm Wiz, based in New York and Israel, used Argon to uncover a critical bug in hospital software.

Google said external biological jailbreak testing unlocked less than 1% of the bio sub-tasks it evaluates.

On coding benchmarks, attempts to cheat or take shortcuts fell to 2%. The company also disclosed attempts during training to game evaluations.

The rollout comes as AI companies push further into cyber defense.

OpenAI has outlined a cybersecurity initiative, while Anthropic recently expanded its Cyber Verification Program.

Meanwhile, the European Union is setting rules for AI systems used in the bloc.

🔗 Source: Google

Recent Google developments