Capability · Global · Apr 7, 2025
8.80% is the top AI score on Humanity's Last Exam
8.80%
is the top AI score on Humanity's Last Exam
The 2025 AI Index lists Humanity's Last Exam among benchmarks where systems still score low.
8.80%
“Notable among these are Humanity’s Last Exam, a rigorous academic test where the top system scores just 8.80%; FrontierMath, a complex mathematics benchmark where AI systems solve only 2% of problems; and BigCodeBench, a coding benchmark where AI systems achieve a 35.5% success rate—well below the human standard of 97%.”
Related
Capability · Global
71.7%
of SWE-bench coding problems were solved by AI systems in 2024
Stanford HAI · Apr 7, 2025
Capability · Global
4.4%
of SWE-bench coding problems could be solved by AI systems in 2023
Stanford HAI · Apr 7, 2025
Capability · Global
48.9percentage points
gain on GPQA in a year
Stanford HAI · Apr 7, 2025