Google has unveiled Gemini 4 Carbon, a new frontier AI model focused on coding, agentic tasks and cybersecurity. Google says Carbon ranks first on Artificial Analysis Intelligence Index 5.0 with a score of 59 and reports strong cybersecurity benchmark results. The model is initially available to selected defenders through Google’s Fairwind Program as safeguards undergo further testing.

Google has unveiled Gemini 4 Carbon, its latest frontier AI model, two weeks after introducing Gemini 4 Argon. The launch steps up competition with OpenAI and Anthropic in coding and cybersecurity systems. Google says Carbon ranks first on version 5.0 of the Artificial Analysis Intelligence Index with a score of 59.

Gains in coding and agentic tasks

Google reports significant gains in coding, agentic tasks and cybersecurity defence. It says it used feedback on the previous launch to improve Carbon in coding, agentic and scientific work, as well as full-stack web development and multimodal understanding. On WebDev Arena, the model gained 89 Elo points to reach fourth place with 1769. The company says Carbon is built for complex, long-running engineering and coding tasks.

Cybersecurity benchmarks