#Knowledge

A 2,500-question exam at the frontier of human knowledge, written by nearly 1,000 subject experts and organised by the Center for AI Safety and Scale AI.

A 57-subject multiple-choice exam, from elementary maths to law and medicine, that became the standard test of an LLM's broad knowledge before frontier models saturated it.

A harder, reasoning-heavy successor to MMLU with about 12,000 questions across 14 disciplines and ten answer options instead of four.

OpenAI's 4,326 short fact-seeking questions with single, indisputable answers, measuring whether models answer correctly or admit they don't know instead of hallucinating.