Skip to content

AIME (American Invitational Mathematics Examination)

← Back to all benchmarks

Date: 2025-03-13

Name: AIME American Invitational Mathematics Examination

Domain: Mathematics

Focus: Pre-college advanced problem solving

Task Types: Problem solving

Metrics: Accuracy

Models: unknown

AI/ML Motif: Reasoning & Generalization

Resources

Benchmark: Visit
Datasets: AoPS website

Keywords

Citation

  • TBD. Aime. March 2025. [Online accessed 2025-06-24]. URL: https://www.vals.ai/benchmarks/aime-2025-03-13.
@misc{www-aime,
  author = {TBD},
  title = {AIME},
  url = {https://www.vals.ai/benchmarks/aime-2025-03-13},
  month = mar,
  year = 2025,
  note = {[Online accessed 2025-06-24]}
}

Ratings

CategoryRating
Software
0.00
No code available
Specification
3.00
Task and Inputs/Outputs are well specified. No system constraints or dataset format is mentioned
Dataset
4.00
Easily accessible data with problems and solutions, but no splits
Metrics
4.00
Correctness is measured, but no grading guidelines are provided.
Reference Solution
0.00
Not given. Human performance stats exist, but no mentions of AI performance
Documentation
3.00
Some background and other information is provided, but it is not comprehensive. No info on how to run an evaluation
Average rating: 2.33/5

Radar plot

AIME (American Invitational Mathematics Examination) radar

Edit: edit this entry