OpenAI has published a write-up, a PDF paper, and a Lean formalization of a result that shows the three-dimensional Navier–Stokes equations can develop a singularity in finite time. The company states the result establishes statements “C” and “D” in the official Clay Mathematics Institute formulation of the Millennium Prize problem. It also states it will not claim the $1 million prize.
The effort used an internal system the lab describes as more capable than GPT-6 Astra. Roughly 10,000 concurrent agents worked the problem across about 88 hours, generating on the order of 2.7 million messages and 130 billion output tokens, according to the company’s public account and contemporaneous reporting.
Key takeaways
- OpenAI published an analytical proof and Lean formalization of a finite-time singularity for 3D Navier–Stokes.
- The company says it will not claim the Clay Millennium Prize.
- Scale of the run: roughly 10,000 agents over about 88 hours.
What the published formalization actually provides
A Lean formalization can be machine-checked independently. That is the concrete deliverable available to outside mathematicians. Outside researchers, including NYU’s Tristan Buckmaster, have raised questions about the research path and whether unpublished work influenced the direction. OpenAI has denied searching user data for the result.
For the broader stack of tools and models that creators evaluate, see How AI is Reshaping the Content Creator Stack in 2026.
What to watch next
Watch independent verification of the Lean formalization. Watch whether the Clay Institute or independent committees issue any formal response. The publication is a research claim with public artifacts, not a prize award.