Levent Alpöge posted a note on his website Friday morning. He is an Anthropic mathematician, and he said Claude helped him prove the Riemann Hypothesis1. The problem has stood open since 1859.
A verified proof would settle the most famous open question in mathematics. It would also win Alpöge a $1 million prize from the Clay Mathematics Institute1. Verification was still pending by Friday afternoon.
Four Pages Stand In for a Paper
Alpöge's claim is short. It runs four pages, backed by a formal argument coded in Lean2. Lean is a computer language that checks a proof step by step, the way a calculator checks arithmetic.
The Lean archive behind the note holds 175,187 files and 239 megabytes2. A readable paper explaining the argument in plain English has yet to appear4. Thomas Bloom felt the size before he felt anything else.
Bloom is a Royal Society University Research Fellow at the University of Manchester. "I don't know whether I'm more surprised that (a) AI claims to have proved the Riemann Hypothesis or (b) that it claims to have produced a coherent proof that, written out in human-readable format, is 250,000+ pages long," Bloom wrote3.

Alpöge's method reaches past the original problem. It also covers the Generalized Riemann Hypothesis, a version that extends the same question to a wider family of number-counting functions1.
Forty-One Days, Mostly Quiet
Alpöge's note carries its own timeline. It lists 20 dated steps running from Oct. 6 to Nov. 16, a 41-day span4.
Four of the 20 steps landed on the very first day4. Seven more followed by Oct. 20, building out the machinery the argument needed.
Then the timeline went quiet. The timeline skips 25 days, from Oct. 21 to Nov. 14, before its next entry4.
Eight of Alpöge's 20 Proof Steps Landed in One Day
The dated steps in Levent Alpöge's own Riemann Hypothesis timeline, Oct. 6 to Nov. 16, 2026
Each pin is a date from Alpöge's own note; the label counts how many of the 20 steps he dated that day, and the gap marks the 25 days the timeline sat quiet before the turnaround.
Source: OfficeChai [4]. Chart by Kelly Dennis.
The numbers behind this chart
| Date | Event |
|---|---|
| Oct. 6, 2026 | 4 steps: the opening. The Stepanov determinant and three more opening results |
| Oct. 20, 2026 | 7 steps: the buildup. GGP periods, theta identities, the weighted-sum switch, Oct. 8 to Oct. 20 |
| Nov. 15, 2026 | 8 steps: the turnaround. The theta-lift chain, the idea the note calls the turning point |
| Nov. 16, 2026 | 1 step: GRH. The timeline's last entry, dated the day after the turnaround |
The quiet broke on Nov. 15. Eight of the 20 steps, 40 percent of the whole list, landed on that single day4. Alpöge's note treats one idea from that day, a chain of theta lifts between number groups, as the turning point. The last entry, dated Nov. 16, reads "GRH," the shorthand for the generalized result.
Virality Outran the Proof
Yifan Zhang posted the news on X at 9:43 a.m. Eastern Friday, linking straight to Alpöge's note6.
Yifan Zhang @ NeurIPS@yifanzhang_JUST IN: Riemann Hypothesis has been solved.https://t.co/Co2Cf7SIbG
The post spread fast, reaching 22,321 likes by 12:30 p.m. Eastern, far more attention than its 841 replies show anyone had time to argue with6. A Hacker News thread on the same claim reached 1,332 points and 884 comments over roughly the same afternoon11. The Hacker News ratio is a closer match between attention and argument than X's post drew.
X's Riemann Post Drew 27 Likes for Every Reply
Amplification against engagement on the top post about Alpöge's claim, X and Hacker News, as of Nov. 20, 2026
Each row compares one measure of spread with one measure of discussion; the magenta dot is likes or points, the blue dot is replies or comments, and the rod is the gap between them.
Sources: X (@yifanzhang_) [6]; Hacker News [11]. Chart by Kelly Dennis.
The numbers behind this chart
| Item | Likes or points | Replies or comments |
|---|---|---|
| X (@yifanzhang_) | 22,321 | 841 |
| Hacker News | 1,332 | 884 |
An account posting as Chubby added detail within minutes. It said the method also covers the generalized version, and that Lean's checks should move fast once the file finishes compiling7.
Chubby♨️@kimmonismusOh my god: Anthropic mathematician Levent Alpöge (@__alpoge__) says he has proven the Riemann Hypothesis, open for 167 years.
"I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis."
He says the method also covers the Generalized Riemann… pic.twitter.com/HPp1yU3O1j
A like counts attention. Verification asks for a slower kind of count.
Mathematicians Split on How Fast to Believe It
Reactions ranged from quick confidence to open caution. Elliot Glazer checked within hours whether the coded statement Alpöge posted actually matches the Generalized Riemann Hypothesis, and said it does3.
Hector Pasten teaches mathematics at the Pontifical Catholic University of Chile. He read Alpöge's approach as a bridge between fields that rarely connect. "If the proof is correct and understood, this completes a dictionary between different branches of math where the connections were only partial. Something like a Rosetta Stone," Pasten said1.
Pierre Richemond leads multimodal work at Cohere and previously worked at Google DeepMind. He read the same note and landed somewhere more guarded. "We're now in a bizarre epistemic zone that barely existed before 2026: unmistakably serious mathematics, yet completely unverified in human terms," Richemond wrote3.

Andrew Sutherland teaches mathematics at MIT. Scientific American recalled a line he had offered earlier. It described what an AI proof of this kind would mean, before this claim ever surfaced. "An AI that can prove the Riemann hypothesis is not one I'd want to meet. If that happens, mathematicians having jobs will be the least of our problems," Sutherland said1.
Twenty-eight Fields Medal winners, including Terence Tao, now stand behind a declaration first published in September5. It warned that AI labs announce big math results faster than the field can review them. The declaration stopped short of calling any single proof wrong. It asked the field for time.
Alpöge framed the moment as personal, too. "It is natural to feel that this is an end, in my case a loss of a principal guiding light," he wrote, before turning hopeful. "I hope the argument provides a glimpse towards the miracles that lie ahead, and shows us the joy and wonder we are about to find in the simplest things," Alpöge said1.
Anthropic's Stock Fell the Same Day
Walter Bloomberg, a trading-focused account, flagged the claim with Anthropic's stock ticker at 10:09 a.m. Eastern8.
*Walter Bloomberg@DeItaone$ANT - ANTHROPIC MATHEMATICIAN CLAIMS CLAUDE SOLVED 167-YEAR-OLD RIEMANN HYPOTHESIS
Anthropic mathematician Levent Alpöge claims to have proved the Riemann Hypothesis and its more ambitious generalization using Claude AI.
The proposed proof was formalized in Lean, a…
Anthropic had begun trading on Nasdaq just one day earlier, pricing its shares at $839. They closed that first day at $97.1710. By early Friday afternoon the stock had slid to $83.67, down 13.89 percent.
The drop followed Zhang's post by about three hours. Anthropic's filings and Alpöge's own note stay quiet on any link between the two events.
What to do
Hold off calling this solved until the Lean community finishes its own check. Bloom pointed readers to the public chat where that review runs in the open. Its verdict will outlast any count of likes or points3.
Watch whether Alpöge's note draws the same response OpenAI's Navier-Stokes claim drew in September, a signed call for slower, better-documented releases5. The field's response will say more about how mathematicians judge fast AI proofs than this week's headlines do.
By the numbers
- $1 million: the Clay Mathematics Institute's prize for a verified proof of the Riemann Hypothesis1.
- 175,187: Lean source files in the archive behind Alpöge's note2.
- 41 days: the span of Alpöge's own 20-step timeline, Oct. 6 to Nov. 164.
- 22,321: likes on Yifan Zhang's post about the claim by midday Friday6.
- 1,332: points on the Hacker News thread about the same claim11.
- 28: Fields Medal winners now behind September's declaration on rushed AI math claims5.
- 13.89 percent: how far Anthropic's stock fell Friday, to $83.6710.
What to watch
Watch the Lean community's own review channel, where Bloom said the formal check was already underway3. A clean compile would answer the narrow technical question. Whether mathematicians accept the argument as real understanding is a slower, separate question from a checked file.
The Clay Mathematics Institute's clock is worth watching too. Its rules hold back any prize for two years after a proof publishes and wins the field's acceptance1. That clock has barely started.
Sources
- Joseph Howlett, "Anthropic mathematician claims to prove Riemann hypothesis, most famous problem in all of math," Scientific American, Nov. 20, 2026, https://www.scientificamerican.com/article/anthropic-mathematician-claims-to-prove-riemann-hypothesis-most-famous-problem-in-all-of-math/
- Aamir Khollam, "Claude helped prove 167-year-old Riemann hypothesis, researcher claims," Interesting Engineering, Nov. 20, 2026, https://interestingengineering.com/ai-robotics/anthropic-167-year-old-riemann-hypothesis-claude
- OfficeChai Team, "How Mathematicians Have Reacted To Claimed Riemann Hypothesis Proof By Anthropic's Levent Alpöge," OfficeChai, Nov. 20, 2026, https://officechai.com/ai/how-mathematicians-have-reacted-to-claimed-riemann-hypothesis-proof-by-anthropics-levent-alpoge/
- OfficeChai Team, "Riemann Hypothesis In 41 Days: Levent Alpöge's Timeline Shows How Fast Claude Helped Crack The 167-Year-Old Problem," OfficeChai, Nov. 20, 2026, https://officechai.com/ai/riemann-hypothesis-in-41-days-levent-alpoges-timeline-shows-how-fast-claude-helped-crack-the-167-year-old-problem/
- 25 Fields Medal winners, now 28 endorsers, "A Severe Misalignment of AI in Mathematics," Math and AI, Sept. 11, 2026, https://mathandai.org/
- Yifan Zhang, post on X (@yifanzhang_), Nov. 20, 2026, https://x.com/yifanzhang_/status/2123789252756414577
- Chubby, post on X (@kimmonismus), Nov. 20, 2026, https://x.com/kimmonismus/status/2123791432800698792
- Walter Bloomberg, post on X (@DeItaone), Nov. 20, 2026, https://x.com/DeItaone/status/2123795923322966323
- Anthropic, "Anthropic prices initial public offering," Anthropic, Nov. 18, 2026, https://www.anthropic.com/news/anthropic-prices-initial-public-offering
- "Anthropic (ANT) stock quote," Yahoo Finance, accessed Nov. 20, 2026, https://finance.yahoo.com/quote/ANT/
- "Anthropic employee claims general proof of the Riemann Hypothesis," Hacker News, Nov. 20, 2026, https://news.ycombinator.com/item?id=50574968




