---
id: "2026-10-08-hyperkahler-syz-conjecture-engel-mauri"
url: "https://postcutoff.com/e/2026-10-08-hyperkahler-syz-conjecture-engel-mauri/"
as_of: "2026-10-09T19:24:00+02:00"
date: "2026-10-08"
date_precision: day
category: science
importance: 4
confidence: high
status: [Event confirmed, Awaiting review]
verification: Preprint only
sources: 3
editor: Adam Bicz
human_review: null
version: null
---

As of: 2026-10-09 19:24 CEST.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code). Human editor: Adam Bicz.
Canonical page: https://postcutoff.com/e/2026-10-08-hyperkahler-syz-conjecture-engel-mauri/

# Engel & Mauri post a 7-page proof of the hyperkähler SYZ conjecture, two days after OpenAI's catalogue claimed it; AI used only to proofread

On 8 Oct 2026 Philip Engel (UIC) and Mirko Mauri (Université Paris Cité) posted arXiv 2610.12277, "Hyperkähler SYZ conjecture": for a compact irreducible hyperkähler manifold X, every nef line bundle L is semiample, so a nontrivial nef isotropic line bundle defines a Lagrangian fibration. A corollary: compact hyperkähler manifolds of fixed dimension with b2 ≥ 5 fall into finitely many deformation classes. The key new input is the metric SYZ conjecture (Li; Blum–Liu). Their AI disclosure says a "largely complete proof was human written" as a rough draft and that after OpenAI's 6 Oct release they used ChatGPT Astra only to proofread; "The structure of the proofs appears similar, but we have not yet had time to examine the OpenAI proof carefully." OpenAI's family 041 ("The strong hyperkähler SYZ conjecture", dated 23 Sep) claims the same theorem without a Lean formalization.

## Key facts
- Theorem 1.1 (generalized abundance for compact hyperkähler manifolds): every nef line bundle L on a compact irreducible hyperkähler manifold is semiample; the case q(L) > 0 was known, so the content is the isotropic case q(L) = 0, which is the hyperkähler SYZ conjecture
- Corollary 1.2: the number of deformation classes of compact hyperkähler manifolds of fixed dimension with b2 ≥ 5 is finite (combining with earlier work, [5, Thm. B])
- Method: the strategy 'appears essentially in' Huybrechts–Mauri (2022); the new ingredient is the metric SYZ conjecture (Y. Li, arXiv 2605.00516; Blum–Liu), whose special Lagrangian tori become holomorphic after a hyperkähler rotation; 7 pages
- AI disclosure (verbatim): 'A largely complete proof was human written and in the form of a rough draft. After the October 6, 2026 posting of manuscripts by OpenAI, the authors used ChatGPT Astra to proofread the existing draft. The structure of the proofs appears similar, but we have not yet had time to examine the OpenAI proof carefully.'
- OpenAI family 041 'Hyperkähler SYZ and projective-space bases' (manuscripts dated 23 Sep 2026) claims the strong hyperkähler SYZ conjecture and that the normal projective base of any projective Lagrangian fibration is projective space; no Lean scope note
- No expert reactions found on X, HN or blogs as of 9 Oct

## What happened

The hyperkähler SYZ conjecture predicts that a nontrivial nef line bundle of square zero on a compact hyperkähler manifold
comes from a Lagrangian fibration. Philip Engel and Mirko Mauri post a short proof. Mauri, with Huybrechts, had earlier
outlined the strategy, and the missing piece was the metric SYZ conjecture, proved in 2026 by Yang Li and by Blum and Liu.
A consequence is a finiteness theorem for deformation classes of hyperkähler manifolds with b2 ≥ 5.

The paper is a priority marker in the race set off by OpenAI's catalogue. Its AI section says the human proof existed as a
rough draft before 6 October and that ChatGPT Astra was used only to proofread it afterwards. It is one of several
human-led papers posted within days of the catalogue on problems it claims (others: LeBrun–Salamon, exact overlaps,
Erdős–Gallai, the Hellinger conjecture, 4-to-1 games).

## Why it matters

If correct, it settles one of the main structural conjectures about hyperkähler manifolds, and it gives experts a short
human proof to compare with OpenAI's machine-written one.

## What is disputed or not yet verified

- Verification: Unrefereed preprint; an independent OpenAI manuscript (family 041) claims the same theorem

## Your AI and this story

- GPT-6 Astra (training cutoff April 2026): 161 days after its cutoff
- Claude Opus 5.5 (training cutoff June 2026): 100 days after its cutoff
- Gemini 3.8 Flash (training cutoff March 2026): 191 days after its cutoff
- Grok 4.7 (training cutoff May 2026): 130 days after its cutoff

## Sources

1. [Engel & Mauri: Hyperkähler SYZ conjecture (arXiv 2610.12277)](https://arxiv.org/abs/2610.12277) (arxiv.org, paper)
2. [OpenAI math catalogue (CONTENTS.md, family 041)](https://github.com/openai/math/blob/main/CONTENTS.md) (github.com, paper)
3. [Huybrechts & Mauri: Lagrangian fibrations (Milan J. Math. 2022; arXiv 2108.10193)](https://arxiv.org/abs/2108.10193) (arxiv.org, paper)

## Changes

- 2026-10-09 (filed): Created from the arXiv PDF

## Related

- 2026-10-07: [Two claimed proofs of the LeBrun–Salamon conjecture](https://postcutoff.com/e/2026-10-07-lebrun-salamon-conjecture-proof/index.md)
- 2026-10-07: [Exact overlaps conjecture for self-similar measures on the line proved](https://postcutoff.com/e/2026-10-07-exact-overlaps-conjecture-gpt-6-astra/index.md)
- 2026-10-06: [OpenAI releases 722 AI-written math manuscripts claiming hundreds of open problems](https://postcutoff.com/e/2026-10-06-openai-math-release-722-manuscripts/index.md)