OpenAI says it has made substantial progress on another Millennium Prize Problem, just days after publishing an AI-generated proposed solution to a forced variant of the Navier–Stokes problem. The company has not revealed which Millennium Prize Problem it is working on, saying it is still deciding how to share the results.
The disclosure adds another chapter to the growing discussion around artificial intelligence and advanced mathematical research. It also comes shortly after OpenAI published its September 8, 2026 research on Navier–Stokes, a result that has attracted attention from mathematicians because of both its technical claims and the circumstances surrounding its development.
OpenAI Confirms Progress but Keeps the Problem Secret
OpenAI made the latest disclosure in a statement provided to The New York Times. The company said that, since completing its Navier–Stokes work, it has made “substantial progress” on another Millennium Prize Problem and is working through how to share the results thoughtfully.
That wording is important. OpenAI has not said that it has solved another Millennium Prize Problem. The company has also not identified the problem, published the underlying research or provided enough technical information for the mathematical result to be independently evaluated.
For now, the confirmed information is limited to OpenAI's statement that it has made substantial progress on another Millennium Prize challenge.
Is OpenAI Working on the Hodge Conjecture?
One of the biggest questions circulating online is whether OpenAI's undisclosed project involves the Hodge Conjecture.
The Hodge Conjecture is one of the seven original Millennium Prize Problems and concerns deep questions in algebraic geometry. It remains an unresolved mathematical problem.
The Hodge Conjecture has attracted attention in online speculation because advanced AI systems are increasingly being used for mathematical reasoning, symbolic work and formal verification. However, there is currently no confirmation from OpenAI that its latest progress concerns Hodge.
Until OpenAI identifies the problem or publishes the underlying research, any connection between the company's latest announcement and the Hodge Conjecture should be treated as speculation rather than established fact.
Editor's note: OpenAI has not identified the second Millennium Prize Problem. References to the Hodge Conjecture are therefore speculation unless and until OpenAI confirms the connection.
What Are the Millennium Prize Problems?
The Millennium Prize Problems were announced by the Clay Mathematics Institute in 2000. Seven exceptionally difficult mathematical problems were selected, with a $1 million prize associated with each successful solution that satisfies the institute's requirements.
The problems span several areas of mathematics, including number theory, computer science, algebraic geometry, mathematical physics, topology and fluid dynamics.
| Millennium Prize Problem | Broad Field | Public Status |
|---|---|---|
| Riemann Hypothesis | Number Theory | Unresolved |
| P versus NP | Computer Science | Unresolved |
| Hodge Conjecture | Algebraic Geometry | Unresolved |
| Birch and Swinnerton-Dyer Conjecture | Number Theory | Unresolved |
| Yang–Mills Existence and Mass Gap | Mathematical Physics | Unresolved |
| Navier–Stokes Existence and Smoothness | Fluid Dynamics | OpenAI has published a proposed result concerning a forced variant; independent review remains important |
| Poincaré Conjecture | Topology | Solved by Grigori Perelman |
The Announcement Comes After OpenAI's Navier–Stokes Research
The latest statement follows OpenAI's September 8, 2026 research publication concerning the Navier–Stokes problem. OpenAI published a write-up of its proposed result together with a formalization in Lean and related research materials.
Navier–Stokes equations are fundamental to the mathematical description of fluid motion. They are used in areas including aircraft design, weather forecasting and the study of fluid flows.
OpenAI said its internal system produced an analytical proof concerning a forced version of the Navier–Stokes problem. In the result described by OpenAI, a smooth external force is applied to the fluid, and the system's proof establishes finite-time singularity formation under that setup.
This distinction matters. The Clay Mathematics Institute's Millennium Prize problem concerns the existence and smoothness of Navier–Stokes solutions in three-dimensional Euclidean space. OpenAI's published work includes an externally forced setup, so it should not be presented as an unqualified statement that the entire original Millennium Prize problem has been independently settled.
OpenAI also said the system produced a formalization of the proof in Lean. According to the company's account, the Lean formalization and verification took an additional 17 hours using GPT-6 Astra.
The research used a system of coordinating AI agents. OpenAI said the group that produced the Navier–Stokes result involved on the order of 10,000 concurrent agents. The company said the agents reached their resolution on September 5, 2026, about 88 hours after the first agents were launched.
OpenAI has also said that it does not intend to claim the $1 million Millennium Prize for this result.
Why the Navier–Stokes Result Has Drawn Scrutiny
The OpenAI result has generated discussion among mathematicians over how the work was produced, how AI systems interacted with the research process and whether outside research could have influenced the result.
NYU mathematician Tristan Buckmaster and Anthropic mathematician Levent Alpöge had been working on related mathematical research. The New York Times reported on the dispute surrounding OpenAI's effort and Buckmaster's concerns about the company's approach and the question of research credit.
OpenAI said that its researchers and agents did not see Buckmaster and Alpöge's work before it was publicly released. The company also said that, following an investigation, Buckmaster's Codex prompts from the preceding two months could not have influenced the internal system, including through training.
OpenAI's published account also says the two groups' mathematical results differed: Alpöge and Buckmaster had produced a result involving external forcing for the Euler equations, while OpenAI's reported Navier–Stokes result involved a different setup.
These questions are separate from the mathematical question of whether OpenAI's published proof meets the exact requirements of the original Millennium Prize problem. Independent mathematical examination remains important.
Clay Institute Says Its Review Process Is Deliberately Unhurried
The Clay Mathematics Institute acknowledged OpenAI's Navier–Stokes announcement in a statement published on September 11, 2026. CMI said its rules govern how proposed solutions are evaluated and how credit is assigned, and described the process as “deliberately unhurried.”
CMI's formal rules are even more specific. Before the institute will consider a proposed solution for a Millennium Prize, three conditions must be satisfied: the solution must be published in a qualifying outlet, at least two years must have passed since publication, and the solution must have received general acceptance in the global mathematics community. CMI also says it does not accept direct submissions of proposed solutions.
That means a mathematical claim involving a Millennium Prize Problem is not transformed into an official Clay-certified solution immediately after publication. The institute's process is designed to allow substantial time for publication, scrutiny and broader mathematical acceptance.
Why the Second Announcement Is Important
The latest disclosure is notable because OpenAI is not simply presenting AI as a tool for routine mathematical calculations. The company is increasingly using advanced AI systems to investigate problems that have challenged mathematicians for decades.
A verified breakthrough on another Millennium Prize Problem would provide another important case study in how AI systems contribute to mathematical research.
At the same time, “substantial progress” can describe different stages of research. It could refer to a new theorem, an important partial result, a promising proof strategy or work that still requires significant mathematical verification.
Without seeing the underlying research, it is not possible to determine which of those possibilities applies to OpenAI's latest claim.
Independent Verification Will Matter
Any eventual Millennium Prize submission would need to withstand detailed mathematical examination.
A result generated with the help of AI is not automatically accepted simply because an AI system produces a convincing argument or a machine-checked component. Researchers still need to establish that the argument addresses the exact mathematical problem and that its assumptions and logical steps are valid.
This distinction is particularly important in the current situation because OpenAI's latest statement about another Millennium Prize Problem provides no technical details about that second project.
The company has said it is deciding how to share the results, so additional technical information could emerge later.
AI Research Is Entering a New Phase
OpenAI's latest announcement comes amid a broader acceleration in AI-assisted mathematical research.
Advanced AI systems are increasingly being used for theorem proving, symbolic reasoning, formal verification, mathematical experimentation and research assistance.
The emergence of multi-agent systems adds another dimension. Instead of asking a single model to solve a problem in one attempt, researchers can deploy multiple AI agents to explore different approaches, test ideas, write code and verify intermediate results.
OpenAI said its Navier–Stokes effort involved millions of agent messages and hundreds of billions of output tokens across its broader experiments. The company reported approximately 2.7 million messages and 130 billion output tokens specifically during the Navier–Stokes effort.
If OpenAI's second Millennium Prize project eventually produces a publicly verifiable mathematical result, researchers will have an opportunity to study not only the mathematics but also the process through which an AI system generated and verified the work.
What We Know So Far
- OpenAI says it has made substantial progress on another Millennium Prize Problem.
- The company has not named the problem.
- OpenAI has not announced a second confirmed solution.
- The Hodge Conjecture is being discussed online, but OpenAI has not confirmed that speculation.
- OpenAI says it is deciding how to share the results.
- The Navier–Stokes work published on September 8 concerns a forced variant and should not be described without that qualification as a completed solution to the original Clay prize criteria.
- Further technical details and independent review will be needed before the significance of the second claim can be assessed.
What Happens Next?
The next major development will likely come when OpenAI identifies the problem and publishes or otherwise shares details of its work.
Until then, the most accurate description is that OpenAI has reported substantial progress, not that it has solved a second Millennium Prize Problem.
For the AI industry, the announcement is another indication that advanced models are being tested against some of mathematics' deepest unanswered questions. For the scientific community, the next stage will be examining the evidence, understanding the methods and determining how much of the claimed progress can be independently verified.
Bottom Line
OpenAI says it has made substantial progress on another Millennium Prize Problem after its recent Navier–Stokes research, but the company has not yet revealed which problem it is or published the underlying results.
The Hodge Conjecture is among the possibilities being discussed online, but there is currently no confirmation from OpenAI linking its latest work to that problem.
OpenAI's September 8 Navier–Stokes publication should also be described carefully: its reported proof concerns a forced variant of the equations, while the Clay Mathematics Institute has its own formal requirements for recognizing a Millennium Prize solution. OpenAI has said it does not intend to claim the $1 million prize for its published result.
Until the company releases more information about the second project, the announcement should be viewed as a report of significant mathematical progress rather than a confirmed second Millennium Prize solution.
CoinAINews will update this report when OpenAI identifies the problem or publishes further technical details.
0 Comments