Humanity's Last Examwith tools

Humanity's Last Exam — extremely hard expert questions across many subjects. “With tools” means the AI is allowed to search the web or run code while answering. Higher is better.

Claude Opus 5.5 by Anthropic holds the best Humanity's Last Exam (with tools) result among tracked models, at 67.7% (released Sep 22 2026). Higher scores are better on this benchmark.

1. Claude Opus 5.5 (Anthropic) — 67.7%; 2. Claude Fable 5.1 (Anthropic) — 65.6%; 3. Claude Fable 5 (Anthropic) — 64.5%; 3. Claude Mythos 5 (Anthropic) — 64.5%; 5. DeepSeek-V4.1-Flash (DeepSeek) — 63.9%.

44 tracked models have a published Humanity's Last Exam (with tools) score. Scores are collected from published lab reports and benchmark sources, with recorded sources credited alongside the results. We do not re-run the evaluations.

GLM-5.3 by Z.ai is the highest-ranked model with downloadable weights on Humanity's Last Exam (with tools), scoring 62.5% at rank 7 overall.