OpenAI has announced that one of its unreleased artificial intelligence models successfully tackled the Navier-Stokes problem, widely regarded as one of the most challenging mathematical puzzles in existence. According to the company, the breakthrough came after just 88 hours of intensive computation, involving 10,000 AI agents working in parallel and consuming millions of dollars in computing resources.
The Navier-Stokes problem is one of the seven Millennium Prize Problems established by the Clay Mathematics Institute, a prestigious collection of mathematical questions that have resisted solution for generations. These problems are considered among the most profound and difficult in the field, with each carrying a $1 million reward for a verified solution.
At its core, the Navier-Stokes problem concerns the equations that describe the motion of fluids such as air and water. These equations have far-reaching applications, from the design of aircraft and weather prediction systems to the study of blood flow in the human body. The specific question posed by the Millennium Prize asks whether these equations can break down over time, a fundamental issue that has puzzled mathematicians for decades.
OpenAI researchers stated that their work demonstrated that the equations can indeed break down over time, thereby providing an answer to the question posed by the prize committee. Mark Chen, OpenAI’s chief research officer, described the achievement as a significant milestone for AI research. He emphasized that the success underscores the potential of artificial intelligence to address some of humanity’s most difficult intellectual challenges.
“This is a significant milestone for AI research, and its promise for the world is that even more of our hardest questions would become possible to answer,” Chen told reporters during a briefing.
Despite the potential $1 million prize attached to the solution, OpenAI has indicated that it will not seek the monetary reward, even if its findings are eventually confirmed by the mathematical community. The company’s focus appears to be on advancing the capabilities of AI rather than claiming financial incentives.
The path to official recognition, however, is not immediate. Under the rules established by the Clay Mathematics Institute, any proposed solution must first be published in a peer-reviewed journal. Following publication, the work must withstand scrutiny from the mathematical community for a period of two years before the institute will convene a committee to formally evaluate the solution.
Professor Martin Bridson, president of the Clay Mathematics Institute, emphasized the importance of a thorough and deliberate review process. “The process of evaluation is deliberately unhurried, and we shall ensure that it is absolutely rigorous,” he told AFP.
OpenAI revealed that it began training the new model in late August, describing it as more capable than any of its previously released systems. The decision to direct the model toward the Millennium Prize Problems came shortly thereafter, following online speculation that a competitor may have solved one of the remaining challenges. At that point, the company instructed the system to tackle all six unsolved Millennium Problems simultaneously.
The Navier-Stokes problem quickly showed signs of progress, prompting OpenAI to allocate substantial computing resources to the effort. According to Sebastien Bubeck, an OpenAI researcher, the final stage of the project involved 10,000 AI agents working on the problem concurrently and communicating with one another in real time.
Chen noted that the computing costs for the project ran “emphatically in the millions of dollars.” Bubeck added that this expenditure was approximately 1,000 times greater than what the company had invested in earlier mathematical achievements, underscoring the scale and complexity of the undertaking.
According to OpenAI’s account, the AI agents reached a solution on Saturday, roughly 88 hours after the project was initiated. The speed and efficiency of the process highlight the transformative potential of advanced AI systems in fields traditionally dominated by human intellect.
While the mathematical community awaits formal verification, the announcement marks a notable moment in the intersection of artificial intelligence and pure mathematics. If confirmed, the solution could pave the way for AI-assisted breakthroughs in other longstanding scientific and mathematical challenges, potentially reshaping the landscape of academic research.