Construction–Verification: A Benchmark for Formalizing Applied Mathematics in Lean 4
Bowen Yang, Yi Yuan, Chenyi Li, Ziyu Wang, Liangqi Li, Bo Zhang, Zhe Li, Zaiwen Wen · 2026
Read Construction–Verification: A Benchmark for Formalizing Applied Mathematics in Lean 4 by Bowen Yang, Yi Yuan, Chenyi Li, Ziyu Wang, Liangqi Li, Bo Zhang, Zhe Li, Zaiwen Wen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-01291.txt.
Description
Bowen Yang and co-authors launch AMBER, a Lean 4 test that forces models to build answers before proving them. Past math tests lean on checking stated claims, skipping crafting concrete values or runnable plans, which applied math prizes most. Their frame compels agents to define plain answers first, then prove them right, across convex study, tuning, number algebra, and high-dimensional chance. Tests show current models strain on build jobs, with broad thought models beating narrow proof specialists, blamed on proof-tuned models losing grip on orders.
Catalog record
| CREATOR | Bowen Yang, Yi Yuan, Chenyi Li, Ziyu Wang, Liangqi Li, Bo Zhang, Zhe Li, Zaiwen Wen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 81776 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602.01291