OpenAI claims abelian Zilber–Pink
Awaiting review
Importance: major (4 of 5)The takeaway
Family 016 of OpenAI’s 6 Oct 2026 release claims the abelian case of the Zilber–Pink conjecture on unlikely intersections: every subvariety of an abelian variety over a number field has only finitely many maximal atypical subvarieties. By a 2022 theorem of Barroero and Dill this would give Zilber–Pink for abelian varieties over any field of characteristic zero. Three more papers claim the curve case in the Siegel threefold A2. Nothing is formalized; the first specialist papers cite it as an ‘announced’ AI-generated proof, and one group says the release pushed them to post their own related work a day later.
Status
- Claim
Awaiting review
- Our reporting
- Medium confidence
- Verification
- Preprint only
- Importance
- Major (4 of 5)
- Last verified
- 29 October 2026
Your AI and this story
- GPT-6 Astra159 days after its cutoff
- Claude Opus 5.598 days after its cutoff
- Gemini 4 Argon189 days after its cutoff
- Grok 4.7128 days after its cutoff
None of these four assistants can know about it. The closest, Claude Opus 5.5, stops 98 days before it.
Key facts
- Manuscripts (OpenAI, all dated 24 Sep 2026, no Lean scope note): ‘The abelian Zilber–Pink conjecture’ (44 pp.), ‘The E×CM component of Zilber–Pink for curves in A2’ (15 pp.), ‘Quaternionic division points on curves in the Siegel threefold’ (46 pp.) and ‘Elliptic squares and Zilber–Pink for curves in A2’ (65 pp.)
- Theorem 1.1 (verbatim): ‘For every number field K, every abelian variety A/K, and every irreducible subvariety X ⊆ A_K̄, there are only finitely many maximal atypical subvarieties of X in S = ⟨X⟩sp.’ The paper adds: ‘No bound on the degrees or heights of the intersection points is assumed, and no simplicity, endomorphism, or complex multiplication hypothesis is imposed on A … it gives no effective or uniform bound’
- Prior work named in the paper: Habegger and Pila proved finiteness for curves in abelian varieties over number fields; Rémond, Viada, Barroero–Kühne–Schmidt and others proved partial results. The proof uses Barroero and Dill’s optimal-subvariety reduction and Dill’s generalized Vojta–Rémond inequality
- Curve case in A2 (abstract of the elliptic-squares paper, verbatim): ‘Combining this result with the companion E×CM and quaternionic-division finiteness theorems, we prove the curve case of Zilber–Pink in A2 over Q̄, without a boundary hypothesis.’
- Barroero, Dill and Kühne (arXiv 2610.10212, 7 Oct, verbatim): ‘Recently, OpenAI has released the preprint [Ope26] claiming a complete proof of Conjecture 1.1 for abelian varieties over Q̄. By [BD22], this would also imply the Zilber-Pink conjecture for abelian varieties over K. This release prompted us to publish a first version of this article the day after on arXiv.’ Their own result settles the abelian analogue of Levin’s question, a case predicted by Zilber–Pink, and was presented at an ICM satellite conference in August 2026
- Netan Dogra and Arnab Saha (arXiv 2610.17325, 12 Oct, verbatim): ‘Recently, an AI-generated proof of the Zilber–Pink conjecture for subvarieties of abelian varieties defined over Q̄ was announced [AI]. By Barroero–Dill this implies the general case of Zilber–Pink for subvarieties of abelian varieties.’ Their paper bounds the degree of points in the Zilber–Pink locus with bounded ramification at a prime, by delta-geometry
- No expert assessment of correctness found by 29 Oct (community notes: only an AI-generated explainer besides the two citing papers)
What happened
“Unlikely intersections” asks when a subvariety meets special subvarieties, such as torsion cosets in an abelian variety, more often than dimension counting predicts. The Zilber–Pink conjecture, formulated by Boris Zilber and Richard Pink in the early 2000s, says that such atypical intersections are controlled by finitely many maximal ones. It contains the Manin–Mumford, Mordell–Lang and André–Oort type statements as special cases. For abelian varieties, finiteness was known for curves (Habegger–Pila) and in several partial ranges.
OpenAI’s manuscript claims the whole abelian case over number fields, in every dimension, through a height-scale analysis built on Barroero and Dill’s reduction and Dill’s generalized Vojta–Rémond inequality. Three further manuscripts claim Zilber–Pink for curves in A2, the moduli space of principally polarized abelian surfaces, a Shimura-variety case beyond abelian varieties.
Why it matters
If correct, this would settle one of the central conjectures in Diophantine geometry for abelian varieties. Because the papers are long, unformalized and so far unread in public by experts, the claim should be treated as unverified.
What is disputed or not yet verified
| Verification | Unrefereed manuscripts; no Lean formalization; cited but not assessed by specialists |
|---|
Reactions
Specialists have cited the claim but not assessed it. Fabrizio Barroero, Gabriel Dill and Lars Kühne say the release prompted them to post their own related paper a day later. Netan Dogra and Arnab Saha describe it as “an AI-generated proof … announced”, and both groups note that Barroero–Dill would turn it into the full abelian Zilber–Pink conjecture. No blog, talk or exposition reviewing the proof was found by 29 Oct.
Sources
7 sources from 2 sites. Numbers match the chips in the text.
7 sources: 6 primary, 1 reaction
Primary
- OpenAI: The abelian Zilber–Pink conjecture (24 Sep)github.com, paper
- OpenAI: Elliptic squares and Zilber–Pink for curves in A2 (24 Sep)github.com, paper
- OpenAI: Quaternionic division points on curves in the Siegel threefold (24 Sep)github.com, paper
- OpenAI: The E×CM component of Zilber–Pink for curves in A2 (24 Sep)github.com, paper
- arXiv 2610.10212: Barroero, Dill, Kühne, Intersecting a curve in an abelian variety with multiples of another curvearxiv.org, paper
- arXiv 2610.17325: Dogra, Saha, Arithmetic jet spaces and the unramified Zilber–Pink conjecturearxiv.org, paper
Reactions
- oai-math-revised: community notes on the abelian Zilber–Pink manuscriptgithub.com, discussion
Changes
- Filed (families 001–125; manuscript introduction and the two citing arXiv papers read)