English

PutnamBench: Evaluating Neural Theorem-Provers on the Putnam Mathematical Competition

Artificial Intelligence 2024-11-05 v2 Computation and Language Machine Learning Logic in Computer Science Programming Languages

Abstract

We present PutnamBench, a new multi-language benchmark for evaluating the ability of neural theorem-provers to solve competition mathematics problems. PutnamBench consists of 1692 hand-constructed formalizations of 640 theorems sourced from the William Lowell Putnam Mathematical Competition, the premier undergraduate-level mathematics competition in North America. All the problems have formalizations in Lean 4 and Isabelle; a substantial subset also has Coq formalizations. PutnamBench requires significant problem-solving ability and proficiency in a broad range of topics taught in undergraduate mathematics courses. We use PutnamBench to evaluate several established neural and symbolic theorem-provers. These approaches can only solve a handful of the PutnamBench problems, establishing the benchmark as a difficult open challenge for research on neural theorem-proving. PutnamBench is available at https://github.com/trishullab/PutnamBench.

Keywords

Cite

@article{arxiv.2407.11214,
  title  = {PutnamBench: Evaluating Neural Theorem-Provers on the Putnam Mathematical Competition},
  author = {George Tsoukalas and Jasper Lee and John Jennings and Jimmy Xin and Michelle Ding and Michael Jennings and Amitayush Thakur and Swarat Chaudhuri},
  journal= {arXiv preprint arXiv:2407.11214},
  year   = {2024}
}

Comments

Accepted at NeurIPS 2024 Datasets & Benchmarks Track

R2 v1 2026-06-28T17:42:14.667Z