x.com/yifanzhang_/status/2097209087129948176

The post overstates what its cited source actually says.

The linked PDF by Tristan Buckmaster does not announce a completed, public proof of the Clay millennium problem. In that statement, Buckmaster says he and Levent Alpöge made public three other results (for incompressible porous media, Boussinesq, and 3D incompressible Euler), and only adds that they "believe" they also have a Navier–Stokes blowup result, which they were not releasing that day because the formal verification was unfinished and there was not yet a presentable writeup. Later in the same document, Buckmaster says OpenAI told him an internal model had a proof of forced Navier–Stokes blowup, but he explicitly says "I have not seen it" and closes conditionally: "If indeed an OpenAI model did close the gap to Navier-Stokes...". That is not evidence that the problem is solved.

Independent of that PDF, Clay Mathematics Institute still marked Navier–Stokes as "Unsolved" on its official millennium-problem pages on September 8, 2026, stating that there is still "no proof for the most basic questions one can ask". Clay’s rules also say a proposed solution is not even considered unless it is published in a qualifying outlet, two years have passed, and it has received general acceptance in the math community.

So the available evidence supports, at most, that there were claims or rumors of a possible proof. It does not support the definite statement that the millennium problem "is solved."