Google's top-ranked AI goes to cyber defenders first — not you

Google announced October 28 that Gemini 4 Carbon, its newest frontier model, ranks first on version 5.0 of the Artificial Analysis Intelligence Index at 59, ahead of Claude Opus 5.5 at 58 and OpenAI's GPT-6 Astra at 57 — two weeks after its predecessor Argon launched.

Carbon's pitch is cybersecurity: Google reports 88.2 percent on SEC-bench Pro for finding real vulnerabilities, 54.8 percent on ExploitGym for turning them into working exploits, and 71.0 percent on CWE-bench for patching them, plus more than double Argon's success rate at generating secure code on internal tests. Google notes the external cyber benchmarks were run without guardrails.

Almost nobody can use it yet: the rollout starts with a subset of cyber defenders through Google's Fairwind Program, with developers, businesses and consumers promised access later and no date given. Introductory pricing matches Argon at $2 per million input tokens and $10 per million output tokens — the model built to break software goes first to the people paid to fix it.