aiqbench analysis
This dataset provides a rigorous benchmark of 50 manually verified, AI-audited theorem statements from the *Annals of Mathematics*, specifically designed to evaluate the limits of current autoformalization tools in Lean 4. By releasing these hard instances without proofs, the authors enable the community to measure genuine progress in mathematical reasoning rather than trivial pattern matching.
A dataset of formalised theorem statements from the Annals of Mathematics
arXiv:2610.18301v1 Announce Type: new Abstract: Recent progress in autoformalisation of mathematics using AI has been impressive. Measuring this progress requires datasets of formal statements that are both hard and trustworthy. We present AnnalsChallenge: 50 formalised theorem statements written by hand in Lean 4 using Mathlib. Each is the main result of a paper published in the Annals of Mathematics since 2020. Every formalisation was manually reviewed and then audited by AI. We release them without formalised proofs.
Abstract:Recent progress in autoformalisation of mathematics using AI has been impressive. Measuring this progress requires datasets of formal statements that are both hard and trustworthy. We present AnnalsChallenge: 50 formalised theorem statements written by hand in Lean 4 using Mathlib. Each is the main result of a paper published in the Annals of Mathematics since 2020. Every formalisation was manually reviewed and then audited by AI. We release them without formalised proofs.
Artificial Intelligence (cs.AI)
arXiv:2610.18301 [cs.AI]
(or arXiv:2610.18301v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2610.18301 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 166 words
Related coverage

Lattice-Based Cryptography Explained
Quantum Zeitgeist · Aug 15, 2026
FaithSieve: Fine-Grained Evaluation of Math Proofs with Faithful Formal Evidence
arXiv · Aug 28, 2026
Learning to Disprove: Formal Counterexample Generation with Large Language Models
arXiv · Oct 15, 2026

Paul Erdos
Quantum Zeitgeist · Aug 21, 2026