OpenAI Says Its AI Cracked Part of a 90-Year-Old Math Problem

OpenAI announced last week that its artificial intelligence made a breakthrough on a famous math problem. The problem is part of the Navier-Stokes equations, which describe how fluids like water and air move. These equations are important for understanding turbulence, a process that is still not fully explained by science. Mathematicians have struggled with parts of this problem for ninety years, and many experts once thought a full solution was still decades away. OpenAI says it used about 10,000 AI agents, working together, to find a solution in just 88 hours.

The Clay Mathematics Institute offers a one million dollar prize for anyone who fully solves the Navier-Stokes problem. It is one of several famous "Millennium Prize" problems in mathematics, and only one of them has ever been solved before. OpenAI's AI agents solved two of the four parts required to win the full prize. The company said the agents exchanged nearly three million messages while working on the problem, testing ideas and checking each other's logic along the way. The whole project used massive computing power, worth about ten million dollars based on normal pricing. OpenAI called the result a "milestone" that shows how fast its newest AI models are improving.

However, the claim quickly caused controversy in the math world. A professor named Tristan Buckmaster said that he and another researcher had also been working on the very same problem for months. He claimed that OpenAI learned about their progress through its own AI tools before announcing its result. OpenAI responded that it never saw the other researchers' work directly and that no user data was misused. The company also said its final proof was different from theirs in several ways. Outside experts have not yet confirmed OpenAI's solution, so the true value of this breakthrough is still unclear, and the dispute may take weeks or months to settle.