Gemini 4 Carbon tops Artificial Analysis Index v5.0, two weeks after Argon

Google released Gemini 4 Carbon just two weeks after Gemini 4 Argon. Sundar Pichai said it ranks #1 on the updated v5.0 Artificial Analysis Intelligence Index at 59. Its cybersecurity numbers are the headline: 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation and 71.0% on CWE-bench for patching. Chrome's security team reports a 10x improvement in patch generation.

Carbon targets long-horizon autonomous engineering, with Google citing significant gains over Argon on extended tasks. Alongside it, DeepMind shipped Genie 4, an interactive world model that now generates real-time simulations with native audio.

Early external signals are positive. Perplexity CEO Aravind Srinivas called Carbon the strongest model he has seen for long-form research, with a new state of the art on the WANDR benchmark, and said it will come to Perplexity Pro and Max once token capacity allows. Cybersecurity is the contested frontier this week: Mistral pitched Large 4 for defensive cyber and OpenAI's Daybreak Red cyber model reached Bedrock GovCloud.

Skeptics on Hacker News questioned whether the Artificial Analysis v5.0 update was coordinated with Google, given Carbon debuted at #1 on a freshly revised index. Release cadence also invites fatigue. Leaks about 'Gemini 4 Xenon' already circulate on r/GoogleGemini. Watch independent evals and API pricing.