Skip to content

MATH-500 Problem Set

A curated subset of 500 problems from the MATH dataset, covering algebra, counting and probability, geometry, intermediate algebra, number theory, prealgebra, and precalculus.

20 models

Higher is better.

  1. 1Ternary Bonsai 2 27BPrism ML98.8
  2. 2Sarvam 105BSarvam98.6
  3. 3QwQ-32BAlibaba98.0
  4. 4Kimi K2Moonshot AI97.4
  5. 5DeepSeek-R1DeepSeek97.3
  6. 6Sarvam 30BSarvam97.0
  7. 7Nemotron Ultra 253BNVIDIA97.0
  8. 8MiniMax M1 80kMiniMax96.8
  9. 9LFM2.5-24B-A2BLiquid AI96.0
  10. 10LongCat-Flash-Lite-SparseMeituan95.8
  11. 11MiniCPM5-2BOpenBMB94.6
  12. 12Magistral MediumMistral94.3
  13. 13MiniCPM5-1BOpenBMB91.6
  14. 14Qwen3 235B-A22BAlibaba91.2
  15. 15DeepSeek V3DeepSeek90.2
  16. 16o1-miniOpenAI90.0
  17. 17LFM2.5-8B-A1BLiquid AI88.8
  18. 18Qwen3 32BAlibaba88.6
  19. 19Claude 3.7 SonnetAnthropic82.2
  20. 20Kanana-2 3B InstructKakao61.2
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Free-form mathematical answers
Tasks
500 problems
Authors
Dan Hendrycks, Collin Burns, Saurav Kadavath, Akul Arora, Steven Basart, Eric Tang, Dawn Song, Jacob Steinhardt (2021)

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.