Google DeepMind announced Gemini 4 Carbon on October 28, calling it the company’s strongest model, with state of the art cybersecurity and long horizon capabilities. Carbon ranks first on version 5.0 of the Artificial Analysis Intelligence Index at 59, and DeepMind said it leads SEC-bench Pro at 88.2 percent for finding vulnerabilities, ExploitGym at 54.8 percent for exploitation, and CWE-bench at 71.0 percent for patching.
The rollout is deliberately narrow. DeepMind said it is engaged in the United States government’s voluntary process for pre-release model testing, and that starting this week Carbon will be available to a subset of cyber defenders, expanding gradually to trusted partners through its Fairwind Program, before any release to developers, businesses and consumers. Pricing matches the introductory pricing of Gemini 4 Argon, launched two weeks earlier, at 2 dollars per million input tokens and 10 dollars per million output tokens, with cached input tokens priced at 95 percent off.
On capability, DeepMind said its internal teams have run Carbon for days at a time on multi-step engineering tasks such as root causing and patching subtle bugs across large codebases, and that it gains 89 Elo points on WebDev Arena to reach fourth place at 1768. On safety measures the company reported that successful cheating on Terminal-Bench 4.0 fell to 0.0 percent, high severity misalignment alerts in live coding dropped by almost 80 percent, and the model resisted every evaluated attack on the Gray Swan indirect prompt injection benchmark, at 0.0 percent against Argon’s 0.7 percent.
The cybersecurity emphasis is the commercial point. DeepMind said the Chrome team found Carbon performed more than ten times better than another lab’s frontier model at generating high quality patches, and that Wiz, its cloud security platform, sees Carbon leading on web penetration testing benchmarks. A model sold first as a defensive security tool, priced like a general frontier model, is Google’s clearest attempt yet to monetise frontier capability through the security budget rather than the chatbot market.