Rumors: OpenAI's "Bel" solved the Navier-Stokes millennium problem

On September 8, a wave of posts surfaced on X claiming that OpenAI had solved the Navier-Stokes equations, one of the Clay Millennium Prize problems in fluid dynamics. As of writing, however, there is no paper or official confirmation — everything remains secondhand rumor.

Confirmed

• Mathematicians Tristan Buckmaster and Levent Alpoge published a paper on a singularity counterexample for the Euler equations, with Terence Tao commenting on Mastodon. This is a real academic event and the direct backdrop to the rumor.
• @Jsevillamol interpreted the work as suggesting Levent and Tristan Buckmaster may have a viable strategy for constructing a Navier-Stokes counterexample, linking the paper as evidence.
• @FateOfMuffins relayed a viral Reddit thread: an OpenAI model claimed a 100-page blow-up proof, in discussions involving Buckmaster and a possible break of the Navier-Stokes problem.

Unconfirmed

• @imjustnewatai claimed an OpenAI internal model codenamed "Bel" had solved the Navier-Stokes problem and that OpenAI would announce it officially within days.
• @basedjensen said the problem had been solved per chatter in the "underlying gatekeeper circle," with finite-time blow-up solutions reportedly reproduced in multiple labs; another post relayed discussions sparked by Elliot Glazer, saying two major AI labs seemed to have independently solved the problem, while explicitly noting the information was secondhand.
• @applesjimmy also claimed OpenAI had apparently cracked the problem, likewise without official confirmation. None of these claims come with a paper or any OpenAI statement.

Why it matters

The Navier-Stokes equations are one of the seven Millennium Prize problems in mathematics. If truly solved by an AI model, it would be a landmark breakthrough for AI-assisted mathematical research. Part of the current buzz stems from the real progress by Buckmaster's team on the Euler singularity counterexample, putting the intersection of fluid mathematics and AI in the spotlight. But until an official announcement or paper lands, readers should treat "OpenAI has solved it" as an unverified insider rumor.

2026-09-08 ~ 2026-09-08 · 19 related posts

Full story(11 episodes)→

• Episode 1: Terence Tao Proposes Keeping Some Open Math Problems Off-Limits to AI(2026-09-03, 2 posts)
• Episode 2: Rumors Claim Claude Solved Navier-Stokes; Terence Tao Responds(2026-09-05, 9 posts)
• Episode 3: Rumor: Anthropic Has Solved Navier-Stokes with Claude(2026-09-06, 2 posts)
• Episode 4: Terry Tao slams closed labs' theorem-proving as viral marketing(2026-09-06, 2 posts)
• Episode 5: Claude Did Not Solve Navier-Stokes, Clarifies Terence Tao(2026-09-06, 2 posts)
• Episode 6: AI-assisted proof of irrationality sparks professional-vs-amateur feud in math community(2026-09-07, 9 posts)
• Episode 7: Terence Tao's latest AI remarks spark buzz, Gary Marcus calls them 'fire'(2026-09-08, 2 posts)
• Episode 8: Rex Douglass Calls 'Not Real Math' Arguments Goalpost-Shifting in AI Debate(2026-09-08, 5 posts)
• Episode 9: Rumors swirl that Claude may have solved the Navier-Stokes millennium problem(2026-09-08, 3 posts)
• Episode 10: Rumors: OpenAI's "Bel" solved the Navier-Stokes millennium problem(2026-09-08, 19 posts)
• Episode 11: Critics slam OpenAI over Navier-Stokes credit dispute(2026-09-08, 3 posts)

• [source] Buckmaster may crack Navier-Stokes; OpenAI claims 100-page blowup proof — FateOfMuffins · 2026-09-08
• Rumor: OpenAI may have solved Navier-Stokes, a Millennium Prize problem — apples_jimmy · 2026-09-08
• Commentary: Buckmaster may have a workable route to a Navier-Stokes counterexample — Jsevillamol · 2026-09-08
• OpenAI Accused of Training on Employees' In-Progress Codex Work to Scoop Them — zephyr_z9 · 2026-09-08
• [source] Rumor: OpenAI's internal model 'Bel' has solved Navier-Stokes, announcement expected — imjustnewatai · 2026-09-08
• Both AI Labs Reportedly Solved Navier-Stokes Independently Amid Growing Drama — basedjensen · 2026-09-08
• Rumor: Navier-Stokes reportedly solved with finite-time blow-ups reproduced at multiple labs — basedjensen · 2026-09-08
• Report: OpenAI's internal model 'Bel' produced a ~100-page proof for Navier-Stokes — imjustnewatai · 2026-09-08
• Anthropic researcher allegedly dropped from OpenAI math paper amid Codex training suspicions — basedjensen · 2026-09-08
• OpenAI rumored to have solved Navier-Stokes, a $1M Millennium Prize problem — iScienceLuvr · 2026-09-08
• TT weighs in: the method extends all the way to Navier-Stokes — basedjensen · 2026-09-08
• Rumor: OpenAI may have cracked Navier-Stokes progress via private Codex sessions — teortaxesTex · 2026-09-08
• Rumor: OpenAI may have solved Navier-Stokes, one of the seven Millennium Prize Problems — mark_k · 2026-09-08
• [source] Navier-Stokes drama: Buckmaster statement references OpenAI's upcoming Condition C/D result — Singularitarian · 2026-09-08
• Rumor: Anthropic may have sniped OpenAI's Navier-Stokes breakthrough amid Codex session drama — teortaxesTex · 2026-09-08
• Mathematicians report major Navier-Stokes progress, say OpenAI may have fully solved it — sachinmaya1980 · 2026-09-08
• Navier-Stokes priority fight looms, and an AGI giant's reputation may take the hit — teortaxesTex · 2026-09-08
• Rumor: OpenAI allegedly swooped in on leaked Navier-Stokes breakthrough, pressured mathematician over authorship — birchlse · 2026-09-08
• Rumors swirl that OpenAI will announce its internal model solved Navier–Stokes — haider1 · 2026-09-08