Google has released Gemini 4 Carbon, a new frontier model that the company says delivers significant gains in coding, agentic tasks and cybersecurity.
According to The Verge, Google highlighted Carbon’s top position on the Artificial Analysis Intelligence Index v5.0 benchmark in its launch announcement. The report says Claude Opus 5.5 and GPT-6 Astra ranked behind it on that evaluation.
The launch places cybersecurity alongside coding and autonomous task performance as a central part of Google’s pitch for the model. However, the available report does not provide additional benchmark scores, testing methodology or technical details about Carbon’s security capabilities.