Papex

Open problem · Physics · Open

Yang–Mills existence and mass gap

Does quantum Yang–Mills theory exist with a positive mass gap?

Summary

Asks for a mathematically rigorous quantum Yang–Mills theory in four dimensions, with a proof that its lightest particle has positive mass.

AI-assisted claims

Partial result claimed · 17 Nov 2026

From Four-Dimensional Wilson Lattice Gauge Theory to Quantum Yang--Mills Theory, with a Conditional Mass-Gap Theorem

“For the mass gap we prove two logical boundaries: an explicit model satisfies all registered inputs with clustering, yet its Hamiltonian has no spectral gap; and a mass gap is equivalent to a uniform exponential rate in the clustering axiom, with the bridge from the volume-uniform transfer-matrix condition to the continuum gap proved.”

By
Zeng Chunping
AI involvement
Text declaration: A mix of human-written and AI-generated. Human-understanding declaration: All parts of this contribution have been checked and understood by a human.
Model
Lean, as stated
Where
Hexagon
Source last read
17 Nov 2026

Note: The source states the construction holds under an explicitly registered package of renormalization-group and analytic assumptions, and that a uniform mass gap requires new infrared input. The model field lists 'Lean', which is a proof assistant, not an AI model.

Partial result claimed · 26 Oct 2026

Existence of quantum chromodynamics and completions of the Standard Model

“As a subproblem, this work also includes a proof of the existence of pure Yang–Mills theory, which is part of the Quantum Yang–Mills Theory Clay Mathematics Institute Millennium Prize problem.”

By
OpenAI
AI involvement
Our models produced a natural language proof for Yang–Mills existence on September 9th, 2026, and then subsequently produced a proof for QCD on September 18th, 2026.
Model
internal OpenAI models, as stated
Where
OpenAI announcement
Formal proof
Lean 4, published by the source
Repository
yang-mills-group/OpenAI-YM-20261028 at 3456e5e · read 21 Nov 2026
Statement file
Challenge.lean
Source last read
12 Nov 2026

Note: OpenAI reports an AI-produced existence proof for pure Yang–Mills and QCD in 4D (Oct 2026), shepherded by a group led by David Gross. On 10 Nov 2026 the group released OpenAI's Lean code for pure Yang–Mills with its first assessment. Two public projects are writing human-readable accounts from that code. The informal manuscript and the QCD proof are not released. The mass gap is not addressed.

Checks

1 review · 2 expositions · 2 responses

  • 26 Oct 2026 · Response · AGMAI: “We call on OpenAI to make public without delay the natural-language and Lean outputs behind its announced existence proofs for Yang–Mills theory, quantum chromodynamics, and a completion of the Standard Model.” Source
  • 28 Oct 2026 · Response · Shepherding committee: Michael R. Douglas, David J. Gross (Chair), Arthur Jaffe, Igor R. Klebanov, Nikita Nekrasov: “Balaban’s ideas were among the important elements that recently enabled OpenAI to follow the theory all the way to infinite distances and prove the existence of its proper continuum limit. But confinement and the mass gap remain very important open challenges. … Before we release it, we would like to be convinced that it is correct. … Once this validation is done, we will release the Lean proof.” Source
  • 10 Nov 2026 · Review · Yang–Mills group: Michael R. Douglas, David J. Gross (Chair), Arthur Jaffe, Igor R. Klebanov, Nikita Nekrasov, with Ron Nissim and Fred Rajasekaran: OpenAI's Lean submission of 28 October 2026 for pure Yang–Mills only, not the informal manuscript and not QCD. The group reports a clean rebuild, an axiom audit, Comparator runs and its own Lean checks, and a review of the definitions that it calls extensive but not exhaustive. It says it does not yet claim to fully understand the proof. “In our judgment the submission proves that a relativistic quantum field theory satisfying the Wightman axioms can be constructed as a scaling limit of four-dimensional Wilson lattice gauge theory for any simple compact gauge group. It does not prove that the infinite-volume limit is unique or that there is a mass gap.” Source
  • 10 Nov 2026 · Exposition · Humanise Yang-Mills project (repository of Martin Hairer): A crowd-sourced project, worked largely through contributors' AI agents, to write a research monograph with detailed proofs traceable to the Lean source at commit 3456e5e. Stage when read on 12 November 2026: source imported and preliminary dependency mapping under way. The README still lists a full build, an axiom audit and a line-by-line mathematical audit as outstanding; the repository's own audit file reports an axiom check of the 17 release roots, finding standard axioms only, run on a prebuilt artifact rather than a clean rebuild. “This project coordinates a crowd-sourced effort to turn the announced Lean formalisation of four-dimensional Yang–Mills theory into a human-readable proof.” Source
  • 11 Nov 2026 · Exposition · Scott Armstrong (CNRS and Sorbonne Université; Courant Institute, New York University): A map and natural-language outline of the released Lean source at commit 3456e5e, with a first-draft manuscript of over 240 pages written by AI agents under human direction and still being revised. The project calls its outline a rough draft written by machines, to be checked, corrected and rewritten by people; its process notes say the reviewers are AI models and that the process does not establish correctness as a referee's report would. An exposition, not a verification of the proof. “The source is about one million lines of Lean in 9,477 files. No informal proof accompanies it. … This repository is an attempt to make that proof readable.” Source

Partial result claimed · 15 Oct 2026

Yang–Mills LLM experiment

“This repository contains a manuscript [QED3] which, if correct, contains a full construction of quantum electrodynamics in three dimensions (QED3), in infinite volume. … For Yang–Mills in three-dimensions (YM3), we have a manuscript [YMM1] which, if correct, contains a verification of the Osterwalder–Schrader axioms for tr[F²].”

By
Ethan Sussman, Levent Alpöge
AI involvement
This repository contains work-in-progress, written largely by LLM. … Novel mathematical ideas should be attributed to Claude (under the guidance of LA). … The manuscripts cited above are the result of cutting material from literally thousands of pages of slop generated by Claude. … It is then fed to Codex to extract a minimal manuscript.
Model
Claude; Codex, as stated
Where
GitHub repository ethanwsussman/Yang--Mills-LLM-experiment
Source last read
30 Oct 2026

Note: The repository's README calls all its claims "tentative, pending verification". The results stated are in three dimensions (QED3, YM3); on four dimensions it says only that they "appear to" have an existence argument. The mass gap is not addressed.

Claims as their sources state them. Listing a claim is not a check by Papex; the status stays the source’s.

All AI-assisted claims

Why it is open

Yang–Mills theory, introduced by Chen Ning Yang and Robert Mills in 1954, generalises Maxwell's electromagnetism by replacing its symmetry group U(1) with a compact non-abelian group G, such as SU(3); its classical equations describe massless waves moving at the speed of light [1]. Arthur Jaffe and Edward Witten list three properties that the quantum theory of the strong force, quantum chromodynamics (QCD), must have: a mass gap, meaning that every excitation of the vacuum has energy at least some Δ > 0; quark confinement; and chiral symmetry breaking [1]. Experiment and simulations "have given strong encouragement" that it has them, but, they wrote, no convincing theoretical computation, "whether or not mathematically complete", demonstrates any of the three [1].

The prize problem asks for a proof that for any compact simple gauge group G a non-trivial quantum Yang–Mills theory exists on four-dimensional space-time and has a mass gap Δ > 0 [1]. Existence means axiomatic properties at least as strong as those of Wightman or of Osterwalder and Schrader, with the short-distance behaviour that asymptotic freedom predicts; a limit obtained only by a compactness argument does not count unless its properties are also established [1]. Jaffe and Witten wrote that "at present we do not know any non-trivial relativistic field theory" satisfying the Wightman axioms in four dimensions, and that even a construction in finite volume would leave the gap open: "no present ideas point the direction to establish the existence of a mass gap that is uniform in the volume" [1].

The Clay Mathematics Institute (CMI) considers a proposed solution only after publication in a qualifying outlet, a wait of at least two years and general acceptance in the global mathematics community; on 14 October 2026 it listed the problem as unsolved [2].

What is known

Physical evidence (1973–2024). Asymptotic freedom, found in 1973 by David Gross and Frank Wilczek and by David Politzer, means that the quantum field behaves almost classically at short distances but not at long ones [1]. Kenneth Wilson's lattice formulation turns the theory into a well-defined statistical model with an inverse coupling β; in four dimensions its continuum limit is believed to need β → ∞ [1, 3]. CMI describes the gap as "discovered by physicists from experiment and confirmed by computer simulations" [2]. In 2020 Andreas Athenodorou and Michael Teper extrapolated to the continuum the spectrum of glueballs, the particles of the pure SU(3) gauge theory without quarks: the lightest "constitutes the theory's mass gap", at 3.405(21) times the square root of the string tension, about 1.65 GeV, although they call a conversion to physical units strictly "not possible" because pure gauge theory and QCD are different theories [4]. In 2024 the BESIII collaboration reported a state, X(2370), whose mass and spin-parity "are consistent with the predictions of the lightest pseudoscalar glueball" [5].

Rigorous theories in lower dimensions (1970s–2026). Constructive quantum field theory, which builds a theory as a limit of approximations, produced interacting theories satisfying the Wightman axioms in two and three space-time dimensions in the 1970s [1]. The two-dimensional abelian Higgs model of David Brydges, Jürg Fröhlich and Erhard Seiler, for which a mass gap was proved, was in Jaffe and Witten's words "the only complete example of an interacting gauge theory satisfying the axioms" [1]. Two-dimensional Yang–Mills theory has been constructed by several methods [3]; in 2026 Ilya Chevyrev and Hao Shen proved that its measure on a torus is invariant for the associated renormalised Langevin dynamics, a random heat flow [8]. In three dimensions the abelian theories have continuum limits, but in 2024 Ajay Chandra, Chevyrev, Martin Hairer and Shen wrote that constructing the pure Yang–Mills measure, "even in finite volume, is open" [3, 9].

Towards four dimensions (1980s–1990s). Tadeusz Bałaban's renormalisation-group analysis of lattice gauge theories on a torus, in three and then four dimensions, gave bounds uniform in the lattice spacing, known as ultraviolet stability [1, 3]. Jacques Magnen, Vincent Rivasseau and Roland Sénéor obtained a form of ultraviolet stability in four dimensions through a continuum regularisation that breaks gauge invariance, which is restored in a limit [3, 9]. Sourav Chatterjee wrote in 2018 that no continuum limit of a lattice gauge theory had been constructed in any dimension above two in which Wilson loop variables behave non-trivially [3].

Mass gaps on the lattice (1978–2025). At strong coupling (small β), Konrad Osterwalder and Seiler proved the area law for Wilson loops, a criterion for confinement, in any dimension [3], and exponential decay of correlations, the lattice form of a mass gap, usually follows from a high-temperature expansion [7]. Shen, Rongchan Zhu and Xiangchan Zhu proved a strictly positive mass gap for SO(N) and SU(N) lattice theories in any dimension within an explicit strong-coupling range [6]. Weak coupling, where the continuum limit lies, is harder [3, 7]; there Arka Adhikari and Sky Cao proved exponential decay for finite gauge groups, a simplification that they say "lessens the direct relevance to physics" [7].

Recent claims. In February 2025 the International Journal of Topology published a paper by Logan Nye presenting "a novel approach to solving the Yang–Mills existence and mass gap problem": it recasts the theory in terms of quantum circuits, states that the construction satisfies the Wightman axioms, and derives a mass gap from the entanglement spectrum of the vacuum [11]. In August 2026 the physicist V. P. Nair posted a preprint, not peer reviewed, giving "the outlines of a proof" of a mass gap in three dimensions, which is not the four-dimensional prize problem [10]. On 28 October 2026 OpenAI submitted to a group chaired by David Gross a Lean formalization of an AI-produced construction of pure Yang–Mills theory in four dimensions; the group published it on 10 November with its own assessment, which is not peer reviewed [12]. In the group's judgment the submission proves that a quantum field theory satisfying the Wightman axioms "can be constructed as a scaling limit of four-dimensional Wilson lattice gauge theory for any simple compact gauge group", with the infinite-volume limit taken only along subsequences, and "does not prove that the infinite-volume limit is unique or that there is a mass gap"; the group adds: "We do not claim to fully understand the proof yet" [12]. On 11 November 2026 CMI's problem page still listed the problem as unsolved [2].

Main approaches

Lattice renormalisation group. Integrate out short-distance fluctuations scale by scale, keeping control of the effective theory at each step [1, 3]. Strongest evidence: Bałaban's ultraviolet stability in three and four dimensions [1, 3]. Main difficulty: the bounds must still be extended to gauge-invariant observables, and Jaffe and Witten write that "new ideas are needed" for a gap uniform in the volume, which would presumably open the way from the torus to all of space [1].

Correlation decay on the lattice. Prove that correlations between distant plaquettes decay exponentially at every coupling, with a correlation length that grows without bound as β → ∞, a problem Chatterjee states in the language of probability [3]. Strongest evidence: mass gaps at strong coupling [3, 6] and, for finite gauge groups, at weak coupling [7]. Main difficulty: Adhikari and Cao call decay of correlations for a four-dimensional theory with a group such as SU(3) "one of the central questions of lattice gauge theory", and weak coupling is the harder regime [7].

Stochastic quantisation. Realise the Yang–Mills measure as the invariant measure of a gauge-covariant random heat flow [8, 9]. Strongest evidence: the two-dimensional measure is invariant for its dynamics [8], and in three dimensions a Markov process on gauge orbits exists for Yang–Mills–Higgs theories [9]. Main difficulty: the three-dimensional process is defined only up to a possible finite-time blow-up, and its authors "do not even know how to show" that it survives for all times, let alone that it has an invariant measure [9].

Geometry of the space of gauge orbits. In a Hamiltonian picture the kinetic energy is a Laplacian on the space of gauge fields modulo gauge transformations; Isadore Singer suggested proving that this space has positive curvature, which bears on the Laplacian's spectrum [1, 10]. Strongest evidence: Nair's 2026 outline of an eigenvalue inequality in three dimensions [10]. Main difficulty: Nair writes that positivity "can be shown in a neighborhood of flat connections" but "difficulties of regularization prevent a general proof" [10].

Large-N duality. Treat SU(N) gauge theory as a string theory with coupling 1/N, which "might give a clear-cut explanation of the mass gap and confinement" for large N [1]. Strongest evidence: "surprising progress" for certain strongly coupled four-dimensional gauge theories with matter [1]. Main difficulty: Jaffe and Witten wrote that "as of yet there is no effective approach to the gauge theory without fermions" [1].

Open sub-questions

  1. Weak coupling, continuous groups. Do plaquette correlations in four-dimensional SU(2) or SU(3) lattice gauge theory decay exponentially at every β, with a correlation length that tends to infinity as β → ∞ [3, 7]?
  2. Bałaban's bounds for observables. Can the uniform bounds on a four-dimensional torus be extended to expectations of gauge-invariant functions such as Wilson loops [1, 3]?
  3. Three-dimensional dynamics. Does the stochastic Yang–Mills–Higgs flow of Chandra, Chevyrev, Hairer and Shen survive for all times, and does it have an invariant measure, a candidate for the three-dimensional theory [9]?
  4. Non-abelian Higgs model in two dimensions. Can the non-abelian Yang–Mills–Higgs measure be constructed in two dimensions, a step that Jaffe and Witten called a qualitative advance and that was still open in 2024 [1, 9]?
  5. A regularised curvature bound. Does Nair's eigenvalue inequality for the three-dimensional gauge-orbit space survive regularisation [10]?
  6. Identifying a glueball. Is X(2370) the lightest pseudoscalar glueball, given that the lattice value for pure SU(3) gauge theory, about 2.6 GeV, is not a direct QCD prediction [4, 5]?

Work on this in Papex

If you have a testable answer to one of the sub-questions above, you can draft it as a private hypothesis in Papex, or prepare the draft from Claude or ChatGPT with the Papex connector. You review the exact version, sign it and decide whether to publish it; a public record that names this problem is listed on this page. A Papex record shows who signed which version and when. It does not check a proof, and it is not a submission for the Clay Mathematics Institute prize.

Key references

  1. Jaffe, A. & Witten, E. Quantum Yang–Mills theory. Official problem description, Clay Mathematics Institute (hosted PDF created 2006). Source · Checked 14 Oct 2026
  2. Clay Mathematics Institute. Yang-Mills & the Mass Gap (problem page), The Millennium Prize Problems (index) and Rules for the Millennium Prize Problems. Source · Checked 14 Oct 2026
  3. Chatterjee, S. (2019). Yang–Mills for probabilists. In Probability and Analysis in Interacting Physical Systems, Springer Proceedings in Mathematics & Statistics, 1–16. 10.1007/978-3-030-15338-0_1 · Checked 14 Oct 2026
  4. Athenodorou, A. & Teper, M. (2020). The glueball spectrum of SU(3) gauge theory in 3 + 1 dimensions. Journal of High Energy Physics 2020(11), 172. 10.1007/JHEP11(2020)172 · Checked 14 Oct 2026
  5. Ablikim, M. et al. (BESIII Collaboration) (2024). Determination of spin-parity quantum numbers of X(2370) as 0−+ from J/ψ → γK0S K0S η′. Physical Review Letters 132(18), 181901. 10.1103/PhysRevLett.132.181901 · Checked 14 Oct 2026
  6. Shen, H., Zhu, R. & Zhu, X. (2023). A stochastic analysis approach to lattice Yang–Mills at strong coupling. Communications in Mathematical Physics 400(2), 805–851 (published online 2022). 10.1007/s00220-022-04609-1 · Checked 14 Oct 2026
  7. Adhikari, A. & Cao, S. (2025). Correlation decay for finite lattice gauge theories at weak coupling. The Annals of Probability 53(1), 140–174. 10.1214/24-aop1702 · Checked 14 Oct 2026
  8. Chevyrev, I. & Shen, H. (2026). Invariant measure and universality of the 2D Yang–Mills Langevin dynamic. Communications on Pure and Applied Mathematics 79(8), 1973–2102. 10.1002/cpa.70043 · Checked 14 Oct 2026
  9. Chandra, A., Chevyrev, I., Hairer, M. & Shen, H. (2024). Stochastic quantisation of Yang–Mills–Higgs in 3D. Inventiones mathematicae 237(2), 541–696. 10.1007/s00222-024-01264-2 · Checked 14 Oct 2026
  10. Nair, V. P. (2026). Towards a proof of mass gap in 3d Yang–Mills theory. Preprint, arXiv:2608.10133 (version 2, 22 August 2026), not peer reviewed. Source · Checked 14 Oct 2026
  11. Nye, L. (2025). Existence and mass gap in quantum Yang–Mills theory. International Journal of Topology 2(1), 2. 10.3390/ijt2010002 · Checked 14 Oct 2026
  12. Yang–Mills group (2026). Validation of the OpenAI Yang–Mills construction: summary of results and independent validation. Assessment 1, 10 November 2026, GitHub repository yang-mills-group/YM; not peer reviewed. Source · Checked 11 Nov 2026

Page reviewed 11 Nov 2026

Source

Open, as listed by Clay Mathematics Institute · Millennium Prize Problems. Checked 11 Nov 2026.

The status is the source’s. Papex does not decide whether a problem is solved.

Status history

  1. Open · Clay Mathematics Institute · Millennium Prize Problems · Checked 11 Nov 2026
  2. Open · Clay Mathematics Institute · Millennium Prize Problems · Checked 29 Sep 2026

Prize

Clay Mathematics Institute · $1,000,000

Prize rules

Only Clay Mathematics Institute awards this prize, under its own rules. Your Papex record is separate: a signed, dated record of your work, not a prize entry.

Records on this problem

No public record addresses this problem yet.

More in Physics