Mathematician Tristan Buckmaster alleges OpenAI learned of his work with Anthropic's Levent Alpöge on Navier-Stokes and used an internal model to solve it (Joseph Howlett/Scientific American)
ThinkingNews Desk · how this was written
OpenAI disclosed that an internal model, described as far more capable than its GPT-6 Astra system, generated a Navier-Stokes existence-and-smoothness proof after an 88-hour run using roughly 10,000 parallel agents. Mathematician Tristan Buckmaster alleges the model accessed his and Anthropic researcher Levent Alpöge’s unpublished work; OpenAI says it cannot rule out the use of de-identified data but denies any direct use of their prompts.
Written from all 4 reports below, not from any single one.
How it was reported
- TechMeme·Mathematician Tristan Buckmaster alleges OpenAI learned of his work with Anthropic's Levent Alpöge on Navier-Stokes and used an internal model to solve it (Joseph Howlett/Scientific American)
OpenAI claims an internal model, described as “significantly more capable than GPT-6 Astra,” solved the Navier-Stokes Millennium Prize problem using 10,000 concurrent agents over 88 hours. Mathematician Tristan Buckmaster alleges the breakthrough derived from his and Anthropic researcher Levent Alpöge’s work, which OpenAI says it cannot rule out but denies any direct use of their prompts.
- TechMeme·Source: Gimlet Labs raised its $300M round after telling investors OpenAI may spend $100M+ annually on its services; OpenAI says it's not a paying customer yet (The Information)
Gimlet Labs raised $300 million after telling investors OpenAI could spend over $100 million annually on its services, though OpenAI says it is not yet a paying customer. Mathematician Tristan Buckmaster alleges OpenAI used an internal model, claimed to surpass GPT-6 Astra, to solve the Navier-Stokes Millennium Prize problem with 10,000 concurrent agents in 88 hours, prompting concerns about the use of his and Anthropic researcher Levent Alpöge’s data.
- TechMeme·Analysis: from 2019 to 2025, gains in pretraining compute efficiency came mostly from data improvements rather than model improvements (Dwarkesh Podcast)
OpenAI claims an internal model, described as “significantly more capable than GPT-6 Astra,” solved the Navier-Stokes Millennium Prize problem using 10,000 concurrent agents running for 88 hours. Mathematician Tristan Buckmaster alleges the system leveraged his and Anthropic researcher Levent Alpöge’s unpublished work, though OpenAI says it cannot rule out that de-identified data from their usage may have contributed to model improvements.
- MIT Tech Review·What OpenAI’s latest controversy tells us about the future of math
OpenAI announced that its internal model solved the Navier–Stokes existence and smoothness Millennium Prize Problem, proving the full equations can break down, while NYU mathematician Tristan Buckmaster and Anthropic researcher Levent Alpöge had earlier posted a proof for a simplified version using publicly available models. The announcement sparked accusations that OpenAI leveraged Buckmaster’s and Alpöge’s AI-assisted work without credit, which OpenAI denies, and the company says it will not claim the $1 million prize.
- TechMeme·Over 10 US states have rolled back tech giants' tax breaks, which top $1B/year in some states, as many others propose similar bills amid a data center backlash (Wall Street Journal)
OpenAI announced that an internal model, described as “significantly more capable than GPT-6 Astra,” solved the Navier-Stokes Millennium Prize problem using 10,000 concurrent agents over 88 hours. Mathematician Tristan Buckmaster alleges the model accessed his and Anthropic researcher Levent Alpöge’s unpublished work on the problem, and OpenAI cautioned it cannot rule out that de-identified data from their use contributed to the breakthrough.
- Hacker News·The Navier–Stokes Millennium Prize Problem
OpenAI disclosed that an internal model produced a Navier–Stokes existence and smoothness proof on September 5 after an 88-hour computation using about 130 billion output tokens. NYU professor Tristan Buckmaster and Anthropic researcher Levent Alpöge claimed a simultaneous breakthrough with Claude and Codex, alleging OpenAI accessed their data; OpenAI denied direct use but acknowledged possible indirect influence from de-identified usage.
- The Next Web·OpenAI publishes its Navier-Stokes proof and says it will not claim the Millennium Prize
OpenAI published a paper, PDF, and Lean formalisation showing that a smooth, initially stationary fluid under a continuous external force can develop a finite-time singularity, claiming resolution of statements C and D of the Navier-Stokes Millennium problem while declining the prize. The proof was generated by roughly 10 000 parallel agents over 88 hours, producing 2.7 million messages and about 130 billion tokens; verification took 17 hours with GPT-6 Astra. The Lean code allows anyone to check mathematical correctness, but the Clay Institute’s prize eligibility remains to be decided by the mathematics community.