An artificial intelligence-generated proof tackling one of modern physics' most difficult challenges has received independent validation from a group of leading scientists led by a Nobel laureate, though a critical aspect of the famous problem remains unresolved. The development follows OpenAI's submission of an existence proof regarding the Yang-Mills problem, which addresses the mathematical framework of the strong force holding the building blocks of matter together.
The assessment was published by the Yang-Mills Working Group, chaired by 2004 Nobel Prize winner in Physics David Gross. The group evaluated a Lean-coded submission provided by OpenAI and made both the complete code and its validation report publicly available. According to OpenAI, its artificial intelligence models generated a written proof on September 9 and finished the Lean formalization process on October 16. Alongside Gross, the assessment team includes physicists Michael Douglas, Arthur Jaffe, Igor Klebanov, and Nikita Nekrasov.
Yang-Mills theory serves as a fundamental mathematical pillar for understanding the strong interaction, which acts as nature's glue by binding quarks together inside protons and neutrons. While physicists have successfully utilized Yang-Mills equations for practical predictions for decades, establishing a rigorous mathematical proof proving that the quantum theory actually exists and possesses its expected properties has long remained elusive. To encourage a complete solution, the Clay Mathematics Institute designated the challenge as one of its seven Millennium Prize Problems in 2000, offering a $1 million reward.
Reviewers stated that OpenAI's submission successfully constructs a relativistic quantum field theory from four-dimensional lattice calculations. The work was formalized using Lean, a proof-assistant language designed to let computers verify the logical steps of a mathematical argument. The verification team reported that its clean rebuild, axiom audit, and other formal checks were successful.
Despite the milestone, the evaluation noted that the accomplishment does not solve the full Millennium Prize problem. Specifically, the submission does not prove the existence of a mass gap, which represents the minimum positive energy separating the vacuum state from the excited states of the theory. Furthermore, the work does not establish that the infinite-volume limit is unique.
The reviewers also included an important caveat in their findings, noting that they utilized OpenAI's Astra tool to assist in examining the code and explicitly stated that they do not claim to fully understand the proof yet.