Math-AI Dataset

AIME25

American Invitational Mathematics Examination 2025.
A compact Hugging Face dataset and data card for the 2025 AIME benchmark, released for mathematical reasoning evaluation and analysis.

Yifan Zhang and Team Math-AI
February 16, 2025  ·  Apache 2.0  ·  Hugging Face Dataset
AIME 2025 Mathematical Reasoning Benchmark Dataset Apache 2.0

Overview

AIME25 provides the 2025 American Invitational Mathematics Examination dataset in a lightweight release format for research use. The accompanying data card documents the dataset location, intended citation, and Apache 2.0 licensing terms.

The release follows the Math-AI AIME dataset convention and is intended for evaluating mathematical reasoning systems on competition-style problems.

2025Exam Year
AIMEBenchmark
HFDataset Host
2.0Apache License

Dataset

Hugging Face

The dataset is hosted at math-ai/aime25.

Repository

The project repository is math-ai-org/aime25.

License

The dataset is released under the Apache License 2.0, allowing use, reproduction, modification, distribution, and sublicensing subject to the license terms.

Data Card

The technical report is included with this site as AIME25.pdf.

Citation

If you use the AIME25 dataset in your research, please cite it as follows:

@misc{aime25,
  title     = {American Invitational Mathematics Examination (AIME) 2025},
  author    = {Zhang, Yifan and Math-AI Team},
  year      = {2025},
  publisher = {HuggingFace},
  url       = {https://huggingface.co/datasets/math-ai/aime25}
}
Open PDF