Google releases Gemini 4 Carbon and says it leads on cybersecurity benchmarks

Google announced Gemini 4 Carbon on October 28 and called it its strongest model, with state-of-the-art cybersecurity and long-horizon capabilities. In the announcement, Koray Kavukcuoglu, SVP at Google DeepMind and Google's Chief AI Architect, said Carbon ranks first, with a score of 59, on the recently updated v5.0 of the Artificial Analysis Intelligence Index. According to The Verge, Claude Opus 5.5 and GPT-6 Astra come in behind it on that index. Google reports top scores on three security benchmarks: 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching. The company also says Carbon writes more secure code on its internal benchmarks, and that in live coding environments with Google's own teams it showed lower rates of misalignment and fabrication. Carbon builds on Gemini 4 Argon, which launched two weeks earlier. Google says it is rolling the model out deliberately and is taking part in the U.S. government's voluntary process for pre-release model testing. Because the model is strong at both finding and exploiting software flaws, it matters to defenders and to anyone worried about misuse. Its release keeps up the fast pace of frontier launches among Google, OpenAI and Anthropic. The scores are Google's own published results.