AIToday
Large Language ModelsAI Business & IndustryThe Verge AIPublished: Oct 6, 2026, 06:00 JST

OpenAI solves Navier-Stokes, but math world is furious

Send on LINE
OpenAI solves Navier-Stokes, but math world is furious

3 Key Points

  1. What happened

    OpenAI said it solved the Navier-Stokes problem, one of seven Millennium Prize Problems worth $1 million each, using an internal model trained from August 28th and 10,000 concurrent agents.

  2. Why it matters

    The result is a historic achievement, yet the field is angry, since mathematicians allege OpenAI rushed its announcement to beat rivals and breached academic norms.

  3. What to watch

    The fallout may hinge on OpenAI's new independent panel of mathematicians, which must help coordinate the release of scores more results its unreleased model produced.

WHO IT HITSUniversity and academic mathematicians, especially those whose unpublished work may have fed OpenAI's models, now face both faster AI-driven results and a debate over credit and norms. Research leaders evaluating AI output will likely face similar disclosure questions.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

The Navier-Stokes result, which OpenAI announced it reached with an internal model trained from August 28th, did not arrive in a vacuum. Over the past year, OpenAI, Anthropic, and other labs have announced breakthroughs on numerous long-standing mathematical problems, in some cases resolving one of the famous Millennium Prize problems. But those results have repeatedly sparked backlash rather than celebration, and OpenAI's response has been to convene a new independent panel of mathematicians to advise it and other AI companies on how new results are presented and released.

That panel, however, arrived abruptly and left many researchers unsure what it will actually do, how much influence it will have, and whether OpenAI will listen. A separate dispute sharpened the tension: mathematician Andreas Thom argued on Mastodon that his and colleagues' interactions with ChatGPT, before OpenAI's triumphant announcement, may have contributed to a result in his area of expertise, non-sofic groups, which OpenAI acknowledged built heavily on his and Gábor Kun's earlier work. Abhishek Saha of Queen Mary University of London framed the broader charge plainly, describing conduct of the kind mathematicians would generally not do.

The coming test, it seems, is less about raw capability than about process. OpenAI says it is learning from earlier mistakes, and the panel has been handed the task of helping coordinate the release of scores more results its unreleased model produced. Whether that wave lands as collaboration or as another reputational crisis may depend on how much the company lets outside mathematicians shape what gets claimed, when, and with whose work acknowledged.

FAQ
How much money comes with solving a Millennium Prize problem?
OpenAI said each of the seven Millennium Prize Problems comes with a $1 million reward for solving it.
What did OpenAI use to solve the Navier-Stokes problem?
OpenAI said it used an internal AI model more powerful than the newly released GPT-6 Astra alongside 10,000 concurrent agents.
What are mathematicians upset about?
Mathematicians have raised allegations of scooping, spying, and flagrant violations of long-standing academic norms, and one, Andreas Thom, accused OpenAI of dishonest behavior about its training data.
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleCHPT hits record September revenue on AI chip test demand