Gemini 4 Carbon tops Artificial Analysis Intelligence Index v5
agents benchmarks claude gemini google
| Source: HN | Original article
Google's Gemini 4 Carbon has risen to the top of the Artificial Analysis Intelligence Index v5.
Google’s Gemini 4 Carbon has topped the latest Artificial Analysis Intelligence Index (AAII) v5, securing the #1 spot with a score of 59. The composite benchmark, which aggregates nine demanding evaluations across knowledge work, coding, science and general knowledge, places Gemini 4 Carbon ahead of all rivals on the intelligence front, though Anthropic’s Claude Opus 5.5 remains the top performer on the related Coding Agent Index.
The ranking follows Google’s rollout of Gemini 4 Carbon earlier this month, a model marketed as its “strongest” with state‑of‑the‑art cybersecurity and long‑horizon capabilities. By climbing to the summit of the AAII, the model demonstrates that Google’s latest architecture can translate its technical promises into measurable performance gains across a broad spectrum of tasks. For enterprises and developers, the result signals a potentially more powerful tool for complex analysis, code generation and scientific inquiry, while also intensifying the competitive race with Anthropic’s Claude Opus 5.5 and OpenAI’s GPT‑6 Astra, both of which sit just behind in the overall intelligence ranking.
Industry watchers will be keen to see how the AAII evolves as providers fine‑tune their offerings and introduce new versions. The next iteration of the index, as well as real‑world adoption metrics for Gemini 4 Carbon—especially in the cybersecurity market where Google initially priced the model at $2 per 1 M input tokens—will indicate whether the benchmark lead translates into broader market traction. As we reported on 29 October 2026, Google’s launch promised a “significant leap” in cybersecurity; the AAII result now provides the first independent validation of that claim.
Sources
Back to AIPULSEN