A Million-Dollar Maths Problem Takes a Big Step — With Help From AI
The Standard Model is the most thoroughly tested theory in all of science. It predicts, with extraordinary precision, what happens when physicists smash particles together at CERN's Large Hadron Collider. Yet underneath that success sits an embarrassing gap: nobody has ever managed to write the theory down as rigorous mathematics and prove that it is sound. Now, one of the hardest pieces of that puzzle may have been solved — by an artificial intelligence.
The problem is called Yang-Mills, and it is one of the seven Millennium Prize Problems set out by the Clay Mathematics Institute in 2000. The full problem carries a $1-million prize. It has two parts. The first is to prove that quantum Yang-Mills theory "exists" — that it can be built up through valid mathematical steps. The second, harder part is to show that a quantity called a "mass gap" emerges from the theory's equations. That second part is still unsolved.
In late October, OpenAI announced that its AI model had solved the first part, the existence problem. Because the proof was far too long and complicated for any human to check by hand, the company asked the eminent theoretical physicist David Gross — a Nobel laureate — to lead a committee of five experts to vet the work before it was released.
On 9 November, the committee gave its verdict: the result "proves the existence of pure Yang-Mills theory according to the problem statements which are standard in physics and mathematics." On the night of 10 November, the committee published a document explaining how the proof works, together with its formal verification in Lean, a programming language that checks every step of a mathematical argument.
The committee is honest about the limits of what it knows. Its members used OpenAI's own AI model to help them digest the code, and they wrote: "We do not claim to fully understand the proof yet." But with the proof and its Lean code now released to the world, mathematicians everywhere can study it, test it and take it apart. As committee member Michael R. Douglas of Harvard University put it: "Despite its centrality to physics, Yang-Mills theory has never been properly understood mathematically, even after more than fifty years of work. No rigorous definition of it existed. Its basic properties, well established in simulations and experiments, were totally out of reach of current mathematics. Not any more."
It is a striking moment for maths: a problem that resisted the world's best minds for half a century has moved forward because a machine produced a proof no human could have written — and other humans, with the help of a computer checker, confirmed it holds up. For students, it is also a reminder that proof is the heart of mathematics: an answer only counts when every step can be checked. If your child would like to get stronger at proof, problem-solving or exam maths, you can find a maths tutor, post a tutoring request and compare tutors, or pick up maths workbooks in the bookshop.
Book your first lesson in maths and build the proof and problem-solving skills that exams — and big ideas — are built on: Book your first lesson
Sources:
- Joseph Howlett, "Scientists confirm OpenAI's blockbuster quantum physics result", Scientific American, 11 November 2026: https://www.scientificamerican.com/article/openais-shepherds-affirm-blockbuster-physics-result/
Curious about maths? Post a request on London Tutors and compare verified maths tutors — online, in person or in a group class.