Google Launches Gemini 4 Carbon for Coding, Agentic Tasks, and Cybersecurity
Google has launched Gemini 4 Carbon, a new frontier model designed to improve AI-assisted coding, agentic tasks, and cybersecurity. According to the company, Carbon achieved the highest score on the Artificial Analysis Intelligence Index v5.0, ahead of Claude Opus 5.5 and GPT-6 Astra.
Carbon builds on Google's Argon model, which was released earlier this month. Its cybersecurity capabilities are currently available only to a select group of cybersecurity defense companies, with broader access planned for other participants in Google's Fairwind Program.
The release positions Gemini 4 Carbon as a model aimed at both general AI capabilities and specialized security applications. Google says its performance reflects the company's ongoing work on frontier models, while the initial restricted rollout suggests that external access will initially be limited to selected partners.