CAM-Bench: A Benchmark for Computational and Applied Mathematics in Lean
CAM-Bench is a new benchmark designed for computational and applied mathematics within the Lean theorem-proving environment. It includes 1,000 proof targets that cover various topics such as optimization and numerical analysis, addressing a gap in existing benchmarks. The benchmark aims to enhance the evaluation of mathematical reasoning in large language models by providing a more comprehensive set of problems.
- ▪CAM-Bench consists of 1,000 proof targets in computational and applied mathematics.
- ▪The benchmark includes problems adapted from textbook exercises and focuses on local definitions and algorithms.
- ▪It aims to complement existing formal mathematics benchmarks by including applied mathematics problems.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.17255 (cs) [Submitted on 17 May 2026] Title:CAM-Bench: A Benchmark for Computational and Applied Mathematics in Lean Authors:Wentao Long, Yunfei Zhang, Chenyi Li, Li Zhou, Chumin Sun, Zaiwen Wen View a PDF of the paper titled CAM-Bench: A Benchmark for Computational and Applied Mathematics in Lean, by Wentao Long and 5 other authors View PDF HTML (experimental) Abstract:Formal theorem-proving benchmarks enable mechanically verifiable evaluation of mathematical reasoning in large language models. However, existing benchmarks mainly focus on Olympiad-style problems and algebraic domains, leaving computational and applied mathematics underrepresented.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
