---
id: "2026-10-26-openai-yang-mills-qcd-existence-claim"
url: "https://postcutoff.com/e/2026-10-26-openai-yang-mills-qcd-existence-claim/"
as_of: "2026-10-27T13:45:00+01:00"
date: "2026-10-26"
date_precision: day
category: science
importance: 5
confidence: medium
status: [Awaiting review]
verification: Lean-verified
sources: 8
editor: human editor
human_review: null
version: null
---

As of: 2026-10-27 13:45 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/2026-10-26-openai-yang-mills-qcd-existence-claim/

# OpenAI claims 4D Yang–Mills and QCD existence

Full title: OpenAI claims a rigorous existence proof for four-dimensional quantum field theories including pure Yang–Mills and QCD; a panel led by David Gross is reviewing it and the AI output is withheld (not the Clay mass gap)

OpenAI's blog post of 26 Oct 2026 ('Working with academic communities to build scientific understanding'; announced by Greg Brockman to ~1.2M views) lists a physics result alongside ETH and Kakeya. According to press accounts, its models produced "the mathematical existence of certain interacting quantum field theories in four dimensions", including QCD and pure Yang–Mills. OpenAI says the Yang–Mills part is "not a solution to the full problem", since the Clay problem also requires a mass gap. The reviewing group (Gross, Douglas, Jaffe, Klebanov, Nekrasov) chose to withhold the AI artifacts and prepare its own write-up. The claim has no public paper, Lean code or expert verdict.

## Key facts
- Claim (as reported by TNW and Superpower Daily from the OpenAI post): a rigorous proof that certain interacting quantum field theories exist in four dimensions, including quantum chromodynamics, and a proof that pure Yang–Mills theory exists. That is the existence part of the Clay Millennium problem, 'not a solution to the full problem'. Andrew Curran's summary (~715k views) lists it as 'Existence of quantum chromodynamics and completions of the Standard Model'
- Reviewers: a group led by David Gross (2004 Nobel Prize in physics) with Michael Douglas, Arthur Jaffe (co-author of the Clay problem statement), Igor Klebanov and Nikita Nekrasov
- Artifacts: the physics group 'chose to withhold the AI-produced artifacts' while it prepares its own write-up. Unlike the ETH and local-smoothing results, nothing is in github.com/openai/math, and the press reports no Lean formalization
- OpenAI on its review arrangements (quoted by Superpower Daily): 'choosing initial shepherds necessarily shapes who gets involved first and creates asymmetry of information.' A 25 Oct AGMAI guest post on Tao's blog had argued against designating mathematicians for AI releases
- Mass gap: AISafetyMemes's widely shared explainer (~185k views) says the result 'proves pure Yang–Mills exists, which is part of' the Millennium problem, but 'the full Clay problem also requires the “mass gap,” so it’s not a clean “AI wins $1M” headline'
- Context: Anthropic's Levent Alpöge and Ethan Sussman have run a public repository since 6 Oct in which Claude tries to finish Bałaban's program; on 16 Oct they reported an unreadable 4D SU(2) Yang–Mills existence 'slop'. OpenAI's 6 Oct release already held 4D operational-QFT papers (scale and conformal symmetry, 26 Sept) and sigma-model papers that NYU's Roland Bauerschmidt said were too hard to verify
- OpenAI blog itself not readable here (HTTP 403); facts are from press relays and X posts
- Engadget (Mariella Moon, 27 Oct) describes the physics result as a proof of 'interacting quantum field theories containing gauge fields and fermions in four dimensions', says the expert group 'has yet to publish anything about it', and gives YANGMILLS4D@gmail.com as the contact for questions
- Date (unverified): a Hacker News commenter quotes the OpenAI post as saying 'Our models produced a natural language proof for Yang–Mills existence on September 9th, 2026'; we could not read the post to confirm
- No public comment from Gross, Douglas, Jaffe, Klebanov or Nekrasov found by 27 Oct

## What happened

OpenAI's 26 Oct post presented three model results with three different review arrangements. The Kakeya and local-smoothing
proof was released as raw output with Lean. The ETH refutation was handed to an open working group. The physics claim went to a
named panel of five senior physicists, who asked OpenAI to keep the AI-produced material private while they write it up
themselves. So, unlike the other two claims, this one cannot be read or checked by anyone outside the panel.

The claim is that interacting four-dimensional quantum field theories, including pure Yang–Mills and QCD, exist as rigorous
mathematical objects. Constructing such a theory in four dimensions has been an open problem in mathematical physics for
decades. It is half of the Clay Millennium problem. The other half, the mass gap, is not claimed.

## Why it matters

If confirmed, it would be the first rigorous construction of a four-dimensional interacting gauge theory, the framework behind
the strong force. The review panel's stature (Gross, Jaffe) means its verdict will carry weight. Until it reports, there is
nothing public to evaluate. Anthropic-side work on the same problem (Alpöge–Sussman) has been public but unreadable since 16 Oct.

## What is disputed or not yet verified

- Verification: Unverified claim; AI artifacts withheld; no paper or Lean code public

## Your AI and this story

- GPT-6 Astra (training cutoff April 2026): 179 days after its cutoff
- Claude Opus 5.5 (training cutoff June 2026): 118 days after its cutoff
- Gemini 4 Argon (training cutoff March 2026): 209 days after its cutoff
- Grok 4.7 (training cutoff May 2026): 148 days after its cutoff

## Sources

1. [OpenAI: Working with academic communities to build scientific understanding (26 Oct 2026)](https://openai.com/index/building-scientific-understanding-with-academic-communities/) (openai.com, official)
2. [TNW: OpenAI says its AI model has disproved the Exponential Time Hypothesis](https://thenextweb.com/news/openai-exponential-time-hypothesis-false-kakeya-yang-mills) (thenextweb.com, press)
3. [Superpower Daily: OpenAI claims three scientific breakthroughs and gives experts different review roles](https://superpowerdaily.com/posts/openai-claims-three-scientific-breakthroughs-and-gives-experts-different-review-roles) (superpowerdaily.com, press)
4. [Engadget: OpenAI announces new discoveries in mathematics, computer science and physics](https://www.engadget.com/2297941/openai-new-discoveries-mathematics-computer-science-physics/) (engadget.com, press)
5. [Greg Brockman on X: 'Announcing three major scientific discoveries…' (~1.2M views)](https://x.com/gdb/status/2114837059462627533) (x.com, discussion)
6. [AISafetyMemes on X: explainer of the three claims (~185k views)](https://x.com/AISafetyMemes/status/2114865886477926790) (x.com, discussion)
7. [Hacker News: Three new discoveries in mathematics, computer science, and physics (969 points)](https://news.ycombinator.com/item?id=50251344) (news.ycombinator.com, discussion)
8. [Andrew Curran on X: summary of the three discoveries (~715k views)](https://x.com/AndrewCurran_/status/2114836071028785289) (x.com, discussion)

## Changes

- 2026-10-27: Engadget details, contact address, HN-quoted 9 Sept date (unverified); no reviewer comment yet
- 2026-10-27 (filed): Created from press relays and X posts (blog post 403; no artifacts public); promoted from the 26 Oct lead

## Related

- 2026-10-27: [Anthropic's Claude also refutes ETH](https://postcutoff.com/e/2026-10-27-anthropic-claude-eth-refutation-3sat-subexponential/index.md)
- 2026-10-26: [OpenAI claims ETH is false](https://postcutoff.com/e/2026-10-26-openai-eth-false-subexponential-sat-claim/index.md)
- 2026-10-26: [OpenAI claims local smoothing and Kakeya in all dimensions](https://postcutoff.com/e/2026-10-26-openai-local-smoothing-kakeya-all-dimensions-claim/index.md)
- 2026-10-16: [Alpöge and Sussman say Claude has an (unreadable) existence argument for 4D Yang–Mills on ℝ⁴, by finishing Bałaban's 1980s program](https://postcutoff.com/e/2026-10-16-yang-mills-claude-balaban-program-alpoge-sussman/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)
- People: [Greg Brockman](https://postcutoff.com/person/greg-brockman/), [Terence Tao](https://postcutoff.com/person/terence-tao/), [Levent Alpöge](https://postcutoff.com/person/levent-alpoge/)
- Archived: 4 posts, listed in https://postcutoff.com/e/2026-10-26-openai-yang-mills-qcd-existence-claim/index.json