# Aaronson: OpenAI math release includes Lean-verified Unique Games proof

> Scott Aaronson says OpenAI's 372 math results include a Lean-verified proof of the Unique Games Conjecture.

*Scott Aaronson says OpenAI's batch of 372 math results contains a machine-checked proof of the Unique Games Conjecture, a claim still awaiting outside verification.*

By Bez · FeaturedDaily
Canonical: https://featureddaily.com/news/aaronson-openai-math-release-includes-lean-verified-unique-games-proof

**What happened:** Scott Aaronson says OpenAI's release of 372 math results includes a Lean-verified proof of the Unique Games Conjecture. He made the claim in a post titled [The Mathocalypse on Shtetl-Optimized](https://scottaaronson.blog/?p=10169), published 8 October.

**The numbers:** The release contains 372 results. According to [AI Weekly](https://aiweekly.co/alerts/openai-says-372-math-results-came-mostly-from-one-prompt), OpenAI says most of them came from one prompt. Four tracked experts shared Aaronson's post within hours.

**The personal angle:** Aaronson notes that his wife, complexity theorist Dana Moshkovitz, worked toward this result for her career.

**Why it matters:** The Unique Games Conjecture is the headline claim in the batch. A Lean-verified proof means a proof assistant has checked it, which is the standard the claim will be judged against.

**The catch:** This is not independently confirmed. The summary of both pages behind this briefing was not read in full, and no independent expert has verified the proof.

**What's next:** Outside mathematicians are expected to check the proof and the Lean formalisation.

## Key takeaways

- OpenAI released 372 math results
- Aaronson says one is a Lean-verified Unique Games proof
- Outside mathematicians have yet to check it

## Sources

- [The Mathocalypse](https://scottaaronson.blog/?p=10169) — Shtetl-Optimized (Scott Aaronson), 2026-10-08
- [OpenAI says 372 math results came mostly from one prompt](https://aiweekly.co/alerts/openai-says-372-math-results-came-mostly-from-one-prompt) — AI Weekly, 2026-10-07