Major AI Benchmarks Explained: What Those AI Scores Actually Measure

Ten major AI/LLM capability benchmarks (MMLU, GPQA, SWE-bench, HumanEval, AIME, etc.) compared on what they measure, format/size, scoring and whether they are saturated — to make sense of the scores models claim.

BenchmarkWhat it measuresFormat / sizeScoringStatus
MMLUKnowledge across 57 subjectsMultiple-choice, ~14,000 (test)Accuracy %Near saturation at frontier
MMLU-ProHarder, more reasoning-heavy MMLU10-choice, 12,032 itemsAccuracy %Still discriminative
GPQA (Diamond)Grad-level bio/phys/chem (Google-proof)4-choice, Diamond 198Accuracy % (25% random)Among hardest knowledge tests
SWE-bench (Verified)Fix real GitHub issuesVerified 500 instancesResolve rate %Not saturated; main SWE bench
HumanEvalPython function generation164 items, run unit testspass@1Saturated; contamination concerns
AIMEHard competition math30 items/yr (answers 0–999)Exact match %Common for reasoning; not saturated
MATHCompetition-level math solving12,500 items, levels 1–5Accuracy %Near saturation
GSM8KGrade-school multi-step word problems~1,000 (test)Exact match %Near saturation
HellaSwagCommonsense sentence completion4-choice, ~10,000 (val)Accuracy %Saturated; losing signal
MMMUCollege-level multimodal (image+text)10,500 (test), 30 subjectsAccuracy %Key multimodal benchmark

Method & sources

Each benchmark's definition, format and size follow its original paper (arXiv) or official repo/site (sources cited per row). ⚠ The "status/saturation" column is qualitative — specific "XX% reached" saturation figures mostly come from secondary aggregators, not the original papers or live official data, so this table omits those percentages and keeps only verifiable qualitative status (e.g. "near saturation", "not yet saturated"). Sizes note the split (e.g. MMLU test ~14,000 items, SWE-bench Verified 500 instances). AIME is borrowed from a math competition with no single official AI-benchmark paper, as noted. Benchmark definitions are relatively stable, though the community releases harder successor versions; refer to official sources.

Source: https://arxiv.org/abs/2009.03300

Retrieved: 2026-07-26