OpenAI’s AI Math Breakthrough Triggers Debate Over What Counts as a Solution

OpenAI says an experimental internal AI model has developed a potential solution to the Navier-Stokes existence and smoothness problem, one of seven Millennium Prize Problems that have remained open for decades.

The company says the research involved around 10,000 AI agents working collectively for about 88 hours. During that period, the agents exchanged roughly 2.7 million messages and generated approximately 130 billion tokens while exploring the problem.

The Navier-Stokes equations are used to describe how fluids such as air and water move. The longstanding mathematical challenge is to determine whether a fluid that begins with a smooth flow can eventually develop a singularity, causing its velocity to become unbounded within a finite amount of time.

OpenAI’s proposed proof argues that this type of breakdown is possible. It describes a fluid vortex becoming increasingly stretched and concentrated until its velocity grows without bound in finite time, even though the system maintains finite total energy.

That mathematical scenario does not mean real fluids can physically reach infinite speeds. Under current scientific understanding, physical effects at sufficiently small scales would become important, while the assumption that fluids behave as continuous substances would no longer accurately describe reality.

Thousands of AI Agents Take on a Historic Problem

OpenAI said the work was conducted using an unreleased internal model that is significantly more capable than GPT-6 Astra, which the company identifies as its most advanced publicly available model.

After the network of AI agents explored different approaches, Astra reportedly spent a further 17 hours formalizing the proposed solution and checking the proof.

If the argument is eventually validated, the implications could extend well beyond theoretical mathematics. Navier-Stokes equations are important to fields including aircraft engineering, weather forecasting and blood-flow research.

A better understanding of the circumstances under which the equations break down could help scientists improve models of highly turbulent and extreme fluid systems.

The approach itself may prove equally important. Coordinating thousands of AI agents to investigate a difficult mathematical question could provide a model for tackling other complex scientific challenges in areas such as medicine, aerospace, energy and materials science.

Researchers Raise Questions About Independence

The claimed breakthrough has also sparked debate over whether OpenAI’s system arrived at the approach independently.

NYU mathematician Tristan Buckmaster and Anthropic researcher Levent Alpöge had been investigating a closely related fluid-dynamics problem and had incorporated AI tools into their research.

Buckmaster said the pair had spent about a year developing an unusual approach involving a smooth external force. He said Codex was used extensively during their work and questioned how OpenAI could have reached a similar direction so soon after learning about their progress.

In posts on Mastodon, Buckmaster described conversations with OpenAI in which he was initially told that the internal research effort had begun without human involvement beyond supplying the mathematical problem.

According to Buckmaster, OpenAI later acknowledged that the initial prompt had been submitted only days earlier, after information about his and Alpöge’s research had reached the company.

OpenAI has disputed the suggestion that its researchers or AI agents had access to the pair’s unpublished work before publication. The company said that no specific user data had been accessed.

However, OpenAI acknowledged that it could not completely eliminate the possibility that de-identified information originating from interactions with its products had contributed to model improvements. It also said the resulting proofs differ substantially from the researchers’ work.

The $1 Million Prize Remains Unclaimed

For now, OpenAI’s proposed argument is not an officially recognized solution to the Navier-Stokes problem.

The Clay Mathematics Institute requires proposed solutions to undergo extensive examination and gain broad acceptance among mathematicians before a Millennium Prize can be awarded.

Independent researchers will therefore need to examine the proof closely and determine whether its reasoning withstands mathematical scrutiny.

OpenAI has said it does not plan to claim the $1 million prize. Instead, the company is presenting the experiment as evidence that increasingly advanced AI systems may be capable of making rapid progress on mathematical problems that have resisted traditional approaches for generations.