OpenAI's Yang-Mills Proof: From Validation to Doubt

Episode 1 · Nobel-Laureate-Led Panel Validates OpenAI's Yang-Mills Existence Proof (2026-11-10, 8 posts)

The Yang-Mills working group, whose scientific committee is chaired by Nobel laureate David Gross, issued its third statement and independent verification report around Nov 10, concluding that OpenAI's Lean-formalized code submitted on Oct 28 (GitHub repo yang-mills-group/OpenAI-YM-20261028) proves the existence of a quantum Yang-Mills theory satisfying the Wightman axioms. This constitutes a partial breakthrough on a million-dollar Millennium Prize problem, though its core—the mass gap—remains unproven. If it holds, it marks a major milestone for LLM-generated proofs in fundamental mathematical physics.

Confirmed

• The working group (chaired by David Gross, with collaboration tied to the Kavli Institute for Theoretical Physics) released its third statement and formal verification report around Nov 10.
• The verified object is OpenAI's Oct 28 Lean-formalized construction of pure Yang-Mills theory existence, repo yang-mills-group/OpenAI-YM-20261028.
• Conclusion: the construction proves existence of a quantum Yang-Mills theory satisfying the Wightman axioms; @IgorCarron relays it as a 4D construction.
• @1411337 and @danintheory note the proof was LLM/AI-generated; @MikePFrank frames it as a partial breakthrough on one of the seven Millennium Prize problems.
• Scientific American reports the review was conducted by an external committee of physicists appointed by OpenAI, with external review mechanisms set up to avoid a repeat of the October controversy.
• Andrew Curran shared news of the report's official release; Reddit users (relayed by @FateOfMuffins) reported the Lean proof as published and verified.

Unconfirmed

• The mass gap component is unproven, leaving an essential gap to a full solution of the Millennium problem.
• Posts in this cluster are mostly relays or second-hand accounts; full report details and final scholarly consensus await primary sources.

Why it matters

• Quantum Yang-Mills existence underpins QCD and the mathematical completion of the Standard Model—something physicists have failed to rigorously establish for decades.
• If an LLM-generated formalized proof survives independent verification, AI moves from auxiliary tool to substantive contributor in frontier mathematics, even as the mass-gap question remains open.

• David Gross-led Yang-Mills group shepherds AI result on QCD existence, statement released — danintheory · 2026-11-10
• Nobel laureate-led committee validates OpenAI's AI proof of quantum Yang-Mills existence — 141_1337 · 2026-11-10
• OpenAI's Yang-Mills construction validation posted by the working group — calabi_and_yau · 2026-11-11
• Yang-Mills Working Group validates OpenAI's Lean construction, but mass gap remains unproven — calabi_and_yau · 2026-11-11
• Reddit: OpenAI's Yang Mills existence Lean proof released and verified on GitHub — FateOfMuffins · 2026-11-11
• Committee Led by Nobel Laureate David Gross Validates OpenAI's 4D Yang-Mills Proof — IgorCarron · 2026-11-11
• OpenAI's 'shepherds' verify AI-generated proof of quantum Yang-Mills existence — scientificamerican · 2026-11-11
• OpenAI's Yang–Mills construction passes Lean verification; mass gap still open — MikePFrank · 2026-11-11

Episode 2 · Experts Question OpenAI's 1000-Page Yang-Mills Proof, Mass Gap Unsolved (2026-11-12, 3 posts)

OpenAI's auto-researchers project released a 1000-page claimed verification of the Yang-Mills existence and mass gap problem, one of the Clay Millennium Prize problems. However, experts' first assessments note that the mass gap remains unproven, with critics arguing the construction may amount to little more than a free theory.

• OpenAI auto-researchers publish Yang-Mills validation, mass gap still unsolved — IgorCarron · 2026-11-12
• Mathematicians call AI's Yang-Mills proof 'slop': no mass gap, object may be trivial — LucaAmb · 2026-11-12
• OpenAI's 1147-page Yang-Mills proof faces its first expert assessment: not proven yet — mishig25 · 2026-11-13