OpenAI publishes its Navier-Stokes proof and says it will not claim the Millennium Prize

Summary

OpenAI has published its paper and machine-checkable Lean proof on the Navier-Stokes equations while choosing not to claim the Millennium Prize. The work required thousands of parallel agents and millions of dollars in compute, but mathematical questions remain regarding external forcing.

Key Takeaways

1. OpenAI has posted a write-up, a PDF paper, and a formalisation of the argument in Lean with a public GitHub repository.
2. “We do not intend to claim the Millennium Prize for this result.”
3. Roughly 10,000 concurrent agents worked the problem across about 88 hours, from 1 to 5 September, consuming 2.7 million messages and around 130 billion output tokens.

Statement Breakdown

• Claimed Facts: 65% of statements the article presents as facts
• Opinions: 25% of statements classified as editorial or subjective
• Claims: 10% of statements surfaced for additional reader evaluation

Credibility & Bias Reasoning

Credibility assessment: The reporting relies on verifiable materials released by OpenAI, including a Lean proof and repository. It offers critical analysis of the technical claims, compute costs, and credit disputes. Key assertions are contextualized against standard mathematical definitions.

Bias assessment: Skeptical Tech Analysis. The article evaluates OpenAI's public announcements with an analytical, slightly skeptical view. It highlights potential gaps between corporate claims and prize criteria without relying on partisan or emotionally charged rhetoric.

Note: Examines corporate technical announcements alongside context on formal mathematical definitions.

Credibility flag: Analytical Reporting

Claimed Facts (5)

• Reports the specific publication and artifact releases by OpenAI.
• Directly reports a metric stated by OpenAI regarding verification time.
• Cites specific operational and compute statistics from the experiment.
• States verifiable facts about recent model releases and internal tooling distinctions.
• Describes verifiable text and credit attributions present on the published page.

Opinions (5)

• Expresses an editorial judgment on which aspect of the release carries the most importance.
• Interprets OpenAI's internal motivations and strategic reasoning.
• Offers an analytical perspective on the central mathematical dispute.
• Presents a subjective evaluation of the cultural impact on mathematical research.
• Draws an inferential conclusion regarding corporate dispute behavior.

Claims (5)

• A corporate claim that is immediately complicated by the decision not to apply for the formal prize.
• Contradicts standard expectations for solving a historic prize problem, raising questions about technical caveats.
• Involves an unsettled intellectual dispute over provenance and unpublished research.
• Introduces a technical condition that may conflict with the unforced formulation of the problem.
• Highlights an unprovable assertion regarding model training data and provenance.

Key Sources

• Ana Maria Constantin — Author, The Next Web
• OpenAI — AI Research Organization
• Tristan Buckmaster — Mathematician