Link to the repository on GitHub: https://github.com/trishullab/PUTNAM
PutnamBench
PutnamBench is a benchmark for the evaluation of theorem-proving algorithms on competition mathematics problems sourced from the William Lowell Putnam Mathematical Competition years 1965 - 2023. Our formalizations currently support three formal languages: Lean 4 $\land$ Isabelle $\land$ Coq. PutnamBench comprises over 1300 manual formalizations, aggregated over all languages.
PutnamBench aims to support research in automated mathematical reasoning by providing a multilingual benchmark for evaluating theorem-proving algorithms. It is released under permissive licenses (Apache 2.0 for Lean 4 and Isabelle, MIT for Coq), and we encourage community contributions (TODO: After initial release)
Author List
George Tsoukalas (corresponding author email: [email protected]) Jasper Lee John Jennings Jimmy Xin Michelle Ding Michael Jennings Amitayush Thakur Swarat Chaudhuri
Statistics
| Language | Count |
|---|---|
| Lean 4 | 514 |
| Isabelle | 514 |
| Coq | 309 |
We also report the number of problems in a certain category. Note that some problems fall under multiple categories.
| Category | Total Quantity |
|---|---|
| Algebra | 218 |
| Analysis | 176 |
| Number Theory | 97 |
| Linear Algebra | 43 |
| Abstract Algebra | 25 |
| Geometry | 22 |
| Combinatorics | 12 |
| Set Theory | 4 |
| Probability | 2 |