Google's Gemini 4 Argon Tops Artificial Analysis HLE-Diamond AI Benchmark

30 Oct 2026, 11:21 am · 3h ago · 1 min read · OfficeChai

Independent AI benchmarking platform Artificial Analysis has introduced HLE-Diamond, a specialized evaluation leaderboard created from a refined 1,000-question subset of Humanity's Last Exam. Designed to evaluate advanced AI reasoning on complex closed-ended research queries across diverse academic disciplines, the benchmark debuted under the platform's v5 release. In initial testing, Google's Gemini 4 Argon model secured the top position, demonstrating superior accuracy and specialized problem-solving capabilities over competing frontier models. The release offers enterprise developers and researchers standardized empirical performance metrics to evaluate state-of-the-art language models on rigorous scientific, mathematical, and multidisciplinary research tasks.