{
  "id": "mmlu",
  "code": "PTL-0187",
  "term": "MMLU",
  "aliases": [
    "Massive Multitask Language Understanding"
  ],
  "category": "benchmarks",
  "definition": "MMLU (Massive Multitask Language Understanding) is a benchmark of four-option multiple-choice questions across 57 subjects, from elementary mathematics to law and medicine, that measures a model's breadth of knowledge and problem solving.",
  "description": "It was the standard headline score for general knowledge for several years and is traditionally reported with five examples in the prompt (5-shot). Frontier models now score near the top, so it is considered saturated, and it has largely been succeeded by harder variants such as MMLU-Pro.",
  "example": null,
  "broader": [
    "benchmark"
  ],
  "narrower": [],
  "related": [
    "mmlu-pro",
    "gpqa",
    "benchmark-saturation",
    "few-shot-prompting"
  ],
  "introduced": 2020,
  "added": "2026-10-10",
  "sources": [
    {
      "title": "Measuring Massive Multitask Language Understanding",
      "authors": "Hendrycks et al.",
      "year": 2020,
      "url": "https://arxiv.org/abs/2009.03300"
    }
  ],
  "url": "https://protologue.com/t/mmlu/",
  "citation": "Protologue. (2026). MMLU. In Protologue: A Taxonomy of Prompting and LLM Techniques (v1.4.0, PTL-0187). https://protologue.com/t/mmlu/"
}