Top AI figures are backing a frontier AI slowdown, while critics warn about regulation, transparency, and open-source risks.
The Clay Mathematics Institute says the Navier-Stokes Millennium Prize Problem has apparently been settled, with a formal review now underway.

The Clay Mathematics Institute says the Navier-Stokes problem has “apparently been settled,” according to the report. The problem is one of the seven Millennium Prize Problems announced in Paris in 2000, each carrying a $1 million prize.
The key point for readers: this is not yet presented as a completed prize decision. A formal review is underway, and CMI describes the process as deliberately unhurried.
The Navier-Stokes problem concerns the equations that govern fluid motion in three-dimensional space. The central question is whether those equations always have a smooth, complete solution.
If the claimed solution withstands review, it would mark a major development for one of mathematics’ highest-profile open problems. For now, the practical takeaway is caution: “apparently settled” signals momentum, not final certification.
CMI says it hopes to see “waves of new human understanding unleashed” as the innovations behind the work are analyzed and interrogated. The institute also notes that new technologies’ increasing ability to accelerate mathematical research has heightened anticipation.
Readers should watch for review updates rather than treating the matter as closed. The strongest signal will be whether the claimed proof survives the institute’s formal process.
The report says the potential solution is accompanied by a bitter dispute. Mathematician Tristan Buckmaster accuses OpenAI of redirecting resources toward the problem after rumors about his research leaked, using his drafts in training data, and rejecting co-author Levent Alpöge, who works at Anthropic, from authorship.
Those are allegations as described in the article, and they add a governance and trust dimension to the mathematical review. The outcome may influence not only the prize process, but also how researchers assess collaboration with AI labs.
Top AI figures are backing a frontier AI slowdown, while critics warn about regulation, transparency, and open-source risks.

AllSpark’s Qwen-based Iris-mini and Iris-pro post leading open-weight search-agent results, according to the paper.

Astra completed 7 of 100 StationeryBench tasks, while MolmoAct2 completed none.

A reported AI-agent campaign hit RubyGems with malicious packages, public-data scraping, and attempted API key theft.