Artificial Analysis Launches HLE-Diamond Benchmark, Gemini 4 Carbon Tops List

Artificial Analysis released its HLE-Diamond ranking based on a refined 1000-question Humanity's Last Exam, with only 13% of models scoring above 50% and Gemini 4 Carbon leading at 69%. Analysis shows open-weight models trail closed-source ones by over 30 points, and higher reasoning effort boosts scores by up to 26 points but costs 11-35x more.

2026-10-30 ~ 2026-10-30 · 4 related posts

• HLE-Diamond Leaderboard: Only 13% of Models Top 50%, Gemini 4 Carbon Leads at 69% — ArtificialAnlys · 2026-10-30
• HLE-Diamond: Leaders Score 16-35 Points Higher on Reasoning, Max Effort Costs Up to 35x — ArtificialAnlys · 2026-10-30
• HLE-Diamond: Claude Opus 5.5 gains 26 points at max effort — at 35x the per-task cost — ArtificialAnlys · 2026-10-30
• HLE-Diamond leaderboard update: open-weights models trail closed by 30+ points — Neurogence · 2026-10-30