Hexagon, a repository for AI-generated mathematics

Hexagon Mathematics FoundationBuilt by: research, non-profit or individualStatus: BetaYear: 2026

Overview

Hexagon is a repository for research in the mathematical sciences and theoretical computer science that accepts results produced with large language models, including results where the contributor claims no human understanding of the work. It opened to submissions on 28 September 2026 and was announced on 6 October 2026 by Benjamin Antieau of Northwestern University, after arXiv told the group that it would not host such results.

Each submission is tagged for how much of the text is AI-generated (primarily AI-generated, a mix, or primarily human-written) and for how much of the work the contributor understands (all, some or no parts). Contributors are the people or organisations responsible for a result, which can include AI labs, and submitters need ORCID-validated accounts and start with a limit of one submission a day. Every version of a work has its own permanent identifier, and Hexagon offers read and write APIs. Formal proofs are not required; contributors can link to Lean formalisations held elsewhere.

Hexagon is run by the Hexagon Mathematics Foundation, a Delaware non-profit with Antieau as its first executive director, and is in public beta. Tech Times reports that it was built with OpenAI’s Codex.

Notes and quotes

  • “Hexagon welcomes submissions by large labs working on language models.” Hexagon’s announcement, quoted by Tech Times.

Sources

This entry was written with the help of AI, from the sources listed. Please check the sources before relying on it, and let us know if anything needs correcting.