---
id: "2026-10-06-openai-de-giorgi-conjecture-dimension-eight"
url: "https://postcutoff.com/e/2026-10-06-openai-de-giorgi-conjecture-dimension-eight/"
as_of: "2026-10-30T13:42:00+01:00"
date: "2026-10-06"
date_precision: day
category: science
importance: 4
confidence: medium
status: [Awaiting review]
verification: Preprint only
sources: 4
editor: human editor
human_review: null
version: "2026-10-28"
---

As of: 2026-10-30 13:42 CET.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/
Canonical page: https://postcutoff.com/e/2026-10-06-openai-de-giorgi-conjecture-dimension-eight/

# OpenAI claims De Giorgi's conjecture in dimension eight

Full title: OpenAI's release claims De Giorgi's 1978 conjecture in its sharp dimension eight, and classifies all stable Allen–Cahn solutions in ℝ⁷ with no energy-growth assumption

"A positive resolution of De Giorgi's conjecture in dimension eight" (OpenAI, 26 Sept 2026, released 6 Oct): Theorem 1.1: if u ∈ C²(ℝ⁸, (−1, 1)) solves Δu = u³ − u with ∂₈u > 0, then u(x) = tanh((e·x − c)/√2); "Neither energy growth nor minimization is assumed". Theorem 1.2: every stable solution v: ℝ⁷ → [−1, 1] is a constant well or a planar heteroclinic, "without an energy-growth assumption". Prior work: Ghoussoub–Gui (n = 2), Ambrosio–Cabré (n = 3), Savin (n ≤ 8 under pointwise limits ±1), del Pino–Kowalczyk–Wei counterexamples (n ≥ 9). Follow-up: Long and Zou, arXiv 2610.28080, "Using the framework developed by OpenAI".

## Key facts
- Claim (abstract, verbatim): 'We resolve De Giorgi's conjecture positively in dimension eight: every entire C² solution u : ℝ⁸ → (−1, 1) of Δu = u³ − u with an everywhere positive directional derivative is a planar heteroclinic. We prove that every stable solution v : ℝ⁷ → [−1, 1] of this equation is a constant well or a planar heteroclinic, without an energy-growth assumption.'
- Why dimension eight is the endpoint: del Pino, Kowalczyk and Wei built non-planar monotone solutions in every dimension ≥ 9 (from non-affine entire minimal graphs, which first exist in ambient dimension 9). Since lower-dimensional solutions extend trivially, dimension 8 implies 4–7, so the claim would settle the conjecture in every dimension where it can hold
- Prior state: Ghoussoub–Gui (1998, n = 2), Ambrosio–Cabré (2000, n = 3), Savin (2009, n ≤ 8 assuming lim_{t→±∞} u(x′, t) = ±1); stable-solution rigidity in ℝ⁷ without growth assumptions was also open (a 6 Oct 2026 preprint by Liu–Luo–Wang–Wei–Wei–Wu needs O(R^{N−1}) energy growth)
- Follow-up: Zhouyu Long and Wenming Zou (arXiv 2610.28080, 20 Oct): 'Using the framework developed by OpenAI, we prove that finite Morse index implies O(R^{n−1}) energy growth for bounded entire solutions … in dimensions 4 ≤ n ≤ 9', giving planar rigidity for 4 ≤ n ≤ 7
- Formal status: no Lean formalization in the release; google-deepmind/formal-conjectures issue #6915 (opened 7 Oct, still open on 28 Oct) proposes marking DeGiorgi_le_eight and DeGiorgi_four…_eight solved on the basis of Theorem 1.1, with an AI-assisted statement comparison
- Reactions: none found from specialists (Savin, del Pino, Wei, Wang) on 28 Oct; community notes (oai-math-revised) list 'Comments and discussion: None found'

## What happened

In 1978 Ennio De Giorgi asked whether the transition layers of the Allen–Cahn equation, a basic model of phase separation, must
be flat when they are monotone in one direction, in dimensions up to eight. The question mirrors the Bernstein problem for minimal
surfaces. It was proved in dimensions 2 and 3. Savin proved it up to dimension 8 under an extra assumption, and it fails from
dimension 9. The full case of dimensions 4–8 without extra assumptions stayed open.

The OpenAI release contains a long manuscript (about 280k characters of text) claiming the sharp case, dimension 8, which implies
all lower ones. It also classifies stable solutions in dimension 7, a separate known open problem. Long and Zou have already used
its framework in a follow-up paper.

## Why it matters

If correct, this closes one of the best-known conjectures in elliptic PDE and phase-transition theory, almost fifty years old.
Our 6 Oct coverage missed it and it surfaced as a lead on 26 Oct. Status: unrefereed AI output, no formal proof, no public expert
assessment yet.

## What is disputed or not yet verified

- Verification: Unrefereed preprint in the OpenAI release; no Lean formalization; used as a framework by one follow-up paper

## Your AI and this story

- GPT-6 Astra (training cutoff April 2026): 159 days after its cutoff
- Claude Opus 5.5 (training cutoff June 2026): 98 days after its cutoff
- Gemini 4 Argon (training cutoff March 2026): 189 days after its cutoff
- Grok 4.7 (training cutoff May 2026): 128 days after its cutoff

## Sources

1. [OpenAI: A positive resolution of De Giorgi's conjecture in dimension eight (26 Sept 2026)](https://github.com/openai/math/blob/main/preprints/De-Giorgis-conjecture-in-dimension-eight-September-26-2026/article.pdf) (github.com, paper)
2. [Long, Zou: On the Allen–Cahn equation (arXiv 2610.28080, builds on OpenAI's framework)](https://arxiv.org/abs/2610.28080) (arxiv.org, paper)
3. [formal-conjectures issue #6915: De Giorgi's conjecture holds in dimension at most 8](https://github.com/google-deepmind/formal-conjectures/issues/6915) (github.com, code)
4. [oai-math-revised: community notes on family 375](https://github.com/mathcommunity/oai-math-revised/blob/main/preprints/De-Giorgis-conjecture-in-dimension-eight-September-26-2026/AGENT_NOTES.md) (github.com, discussion)

## Changes

- 2026-10-28 (filed): Created (science agent; paper, openai/math CONTENTS, oai-math-revised notes, arXiv 2610.28080, GitHub issue)

## Related

- 2026-10-26: [AI-assisted results in the 26 Oct 2026 sweep](https://postcutoff.com/e/2026-10-26-ai-assisted-results-october-26-listing/index.md)
- 2026-10-24: [oai-math-revised: mathematicians launch a community hub to revise OpenAI's AI-written math manuscripts (guest post on Tao's blog)](https://postcutoff.com/e/2026-10-24-oai-math-revised-community-hub/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)