Instruction Following & Long Context
CAIS AI DashboardHumanity's Last Exam
Expert-level questions across mathematics, science, and humanities, as published by the CAIS AI Dashboard.
Metric
Accuracy
Results shown
65
Publisher
CAIS AI Dashboard
Snapshot
Fetched 2026-10-23T00:21:17.611Z
65 of 65 models
65 of 65 models
Higher is better
Op
Opus 5.5Fa
Fable 5.1
Fa
Fable 5Op
Opus 5
Op
Opus 4.8
Op
Opus 4.7Op
Opus 4.6
Op
Opus 4.5So
Sonnet 4.6So
Sonnet 4.5Ha
Haiku 4.5So
Sonnet 4So
Sonnet 3.7List bars visualize relative performance. Plot bars use published 0–100 scores, shrink to a 1.5rem minimum, then scroll horizontally.
Scores are published snapshots and are only comparable within the same benchmark, version, metric, and evaluation configuration.