Gemini 4 Carbon: Our next frontier model

Judgment

Enterprise platform leads and security engineering chiefs must decide whether to re-benchmark agent pipelines against Carbon through the gated Fairwind Program now, or hold on Gemini 4 Argon, which shipped two weeks ago. Because pricing is identical, the real cost is rework against an access timeline Google controls.

Basis

Koray Kavukcuoglu, SVP of Google DeepMind and Chief AI Architect at Google, called Gemini 4 Carbon "our strongest model with state-of-the-art cybersecurity and long horizon capabilities." The numbers support a capability argument, not a price one: Carbon ranks #1 on Artificial Analysis Intelligence Index v5.0 at 59, scores 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation and 71.0% on CWE-bench for patching, and reaches #4 at 1768 on WebDev Arena, a +89 elo jump. Rollout is staged: a subset of cyber defenders this week, then trusted partners via the Fairwind Program, with Google also engaged in the U.S. government's voluntary pre-release testing process. Pricing is the same introductory rate as Argon: $2 per million input tokens and $10 per million output tokens, with cached input at 95% off the input price. Mechanism: equal pricing strips out budget as a tiebreaker, so the decision reduces to whether a team's benchmark gains on long-horizon coding and vulnerability work justify building on a partner-gated model whose broader developer access has no date.