Gemini 4 Carbon Takes #1 on the Artificial Analysis Intelligence Index

Google's new frontier model scores 59 on Artificial Analysis Intelligence Index v5.0, with big claimed gains in cybersecurity. Cyber defenders get it first.

By Gobi Daily2 min readFact-checked

Watch Gobi's Short

On October 28, 2026, Google announced Gemini 4 Carbon, which it calls its strongest model yet, with state-of-the-art cybersecurity and long-horizon capabilities. It comes just two weeks after Gemini 4 Argon.

#1 on the leaderboard

Carbon scores 59 on the newly updated Artificial Analysis Intelligence Index v5.0, the top score on the board. Artificial Analysis lists Claude Opus 5.5 (Max) right behind it at 58. The index, launched on October 27, combines nine evaluations across knowledge work, coding, science and general reasoning, and most of it now runs on private test sets.

Built for cyber defense

Google says Carbon covers the whole defense cycle, from finding bugs to fixing them. On external benchmarks run without guardrails, Google reports:

  • 88.2% on SEC-bench Pro for finding real vulnerabilities
  • 54.8% on ExploitGym for checking which bugs can actually be exploited
  • 71.0% on Collinear’s CWE-bench for writing patches

Google also says Carbon more than doubles Argon’s success rate at generating secure code on internal tests, and that internally it has run for days on hard engineering problems like tracking down subtle bugs across large codebases.

Safety claims

According to Google, Carbon cut successful cheating on Terminal-Bench 4.0 to 0.0%, roughly halved unsupported claims compared with Argon, and resisted every attack in Gray Swan’s indirect prompt injection benchmark, down from a 0.7% attack success rate for Argon.

Who can use it

Not most people, yet. Starting this week, Google is giving Carbon to a subset of cyber defenders and will phase in more trusted partners through its Fairwind Program. It says developers, businesses and consumers will get it as soon as possible. Google is also taking part in the U.S. government’s voluntary pre-release model testing.

When it does launch, the introductory API price matches Argon at $2 per million input tokens and $10 per million output tokens, rising to $4 and $20 after the introductory period.

Sources

  1. Gemini 4 Carbon: our next frontier modelGoogle
  2. Artificial Analysis Intelligence Index v5.0Artificial Analysis

Fact-checked against the primary sources above on Oct 28, 2026. Spot an error?Request a correction.

More in Google →