Fifteen elite high-school competition math problems from 2025, used as a clean test of fresh-year reasoning.
AIME 2025 is the 2025 edition of the qualifying contest for the US national math olympiad. Each problem demands creative reasoning across algebra, geometry, number theory, or combinatorics, and the answer is always an integer between 0 and 999.
Models are asked to solve each problem and produce a single integer. Scoring is percent of correct integers. Leaderboards typically report majority-vote pass@1 across multiple samples to reduce variance on the small problem set.
No scores yet for this benchmark.
Not enough scored models yet.
Not enough scored models yet.
Same format and difficulty. AIME 2025 has been around longer, so almost every modern model has been evaluated on it. AIME 2026 is fresher and a cleaner test of contamination resistance.
Based on score correlations across our database.
