{
  "id": "aime",
  "code": "PTL-0199",
  "term": "AIME",
  "aliases": [
    "AIME 2024",
    "AIME 2025",
    "American Invitational Mathematics Examination"
  ],
  "category": "benchmarks",
  "definition": "AIME, the American Invitational Mathematics Examination, is a US high-school mathematics competition whose problems, each with an integer answer, are widely used as a benchmark for reasoning models, usually identified by year, as in AIME 2024 or AIME 2025.",
  "description": "Each exam has only 15 problems, so scores are typically averaged over many sampled attempts and vary noticeably between runs. A new exam each year provides problems written after many models' training data was collected, which limits contamination.",
  "example": null,
  "broader": [
    "benchmark"
  ],
  "narrower": [],
  "related": [
    "math-benchmark",
    "reasoning-model",
    "test-time-compute-scaling",
    "self-consistency"
  ],
  "introduced": null,
  "added": "2026-10-10",
  "sources": [
    {
      "title": "Learning to reason with LLMs",
      "authors": "OpenAI",
      "year": 2024,
      "url": "https://openai.com/index/learning-to-reason-with-llms/"
    },
    {
      "title": "s1: Simple test-time scaling",
      "authors": "Muennighoff et al.",
      "year": 2025,
      "url": "https://arxiv.org/abs/2501.19393"
    }
  ],
  "url": "https://protologue.com/t/aime/",
  "citation": "Protologue. (2026). AIME. In Protologue: A Taxonomy of Prompting and LLM Techniques (v1.4.0, PTL-0199). https://protologue.com/t/aime/"
}