Claude May Have Just Proved the Riemann Hypothesis. Here's Why You Shouldn't Celebrate Yet
Claude May Have Just Proved the Riemann Hypothesis. Here's Why You Shouldn't Celebrate Yet On Friday, November 20, 2026, a mathematician at Anthropic posted a four-page note on his personal website. The claim inside it
Claude May Have Just Proved the Riemann Hypothesis. Here's Why You Shouldn't Celebrate Yet
On Friday, November 20, 2026, a mathematician at Anthropic posted a four-page note on his personal website. The claim inside it is the kind of sentence that stops you mid-scroll: "I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis."
Levent AlpΓΆge is a number theorist on Anthropic's technical staff, formerly of the Harvard Society of Fellows. The Riemann Hypothesis is the most famous unsolved problem in mathematics β open since 1859, one of the seven Millennium Prize Problems, worth $1 million if it ever gets solved properly. If this claim holds up, it is one of the biggest moments in the history of mathematics, and an AI system helped do it.
That "if" is doing an enormous amount of work, and this article is mostly about the "if". The claim has not been independently verified. No outside expert has reviewed the proof. The man who wrote it says he has not yet mastered all of it himself. So: excitement, yes. Celebration, not yet. Here's what actually happened, what it would mean, and what to watch next.
What the Riemann Hypothesis actually is
Prime numbers β 2, 3, 5, 7, 11, and so on β are the building blocks of arithmetic. They get rarer as numbers grow, but not in a smooth, predictable way. Their distribution has irregularities, and mathematicians have spent generations trying to understand the pattern behind them.
The Riemann Hypothesis, first proposed by Bernhard Riemann in 1859, says something remarkably precise about that pattern. Technically, it says that all the "nontrivial" zeros of the Riemann zeta function lie on a single vertical line in the complex plane, the line with real part equal to one-half. You don't need to follow the complex analysis to get the stakes: a proof would give mathematicians far tighter control over how primes are distributed, and it would remove an assumption from a huge body of mathematical results that currently depend on the hypothesis being true.
AlpΓΆge's claim goes further than the original problem. His note says the method also resolves the Generalized Riemann Hypothesis, which extends the same idea to a broader family of mathematical functions, and he expects the Extended Riemann Hypothesis to follow with more work. That last part is an expectation, not a result.
What AlpΓΆge actually released
Here's the unusual part: there is no conventional paper. Nobody has published a written argument explaining the complete proof, because β by AlpΓΆge's own account β the full written argument would run to more than 250,000 pages, and he has not yet mastered all of the mathematics involved. He plans to write it up properly.
What he released instead is a formalization in Lean 4, the programming language and proof assistant that checks mathematical reasoning step by step, spread across 14 archive parts and reported to run to around 17 million lines of Lean code. That is the whole public record of the proof right now: a four-page note saying "I believe I have discovered a proof," plus a mountain of machine-checked code.
That format is both the strength and the strangeness of this announcement. It means the claim is checkable in a way a 250,000-page manuscript never could be β a computer can verify each logical step. It also means that, as of today, no human being fully understands the argument, including its author.
What Lean verification means, and what it doesn't
Lean is worth understanding, because it's doing a lot of the heavy lifting in how seriously people are taking this. A proof written in Lean's formal language can be checked against explicit definitions and logical rules. Researchers can inspect what the final theorem actually says, and which assumptions it depends on.
That gives the mathematical community a concrete, fast first check: does it compile? Kevin Buzzard, a mathematician at Imperial College London known for his work with Lean, gave the suitably practical response to New Scientist: "Let's see if it compiles first."
But compiling is only the first of several checks, and the later ones are harder:
- Does the theorem actually state the Riemann Hypothesis? A computer can correctly verify a statement that falls short of the problem people intended to solve. The formal statement has to be checked against the real thing.
- What does it depend on? Lean supports standard foundational axioms, so the meaningful check is whether the result relies on acceptable foundations, without extra assumptions or hidden gaps.
- Does anyone understand it? Even if the code checks out, mathematicians need to understand how it works before broadly accepting it β well enough to explain it, teach it, and build on it. Youness Lamzouri, a professor of mathematics at the University of Lorraine, told IBM Think that process could take months or even years: "Mathematics is mainly about understanding why something is true. We need to understand the ideas in the proof well enough to explain them, teach them and build on them to extend human knowledge."
That last point is the one that makes this story different from a normal breakthrough. If the proof survives, mathematicians could have strong formal evidence for a result while still working to understand the ideas that make it true. As IBM's coverage put it, the unusual presentation leaves mathematicians with two tasks: determining whether the computer-checked proof establishes the claimed result, and understanding the mathematical ideas behind it.
Why this claim is being taken seriously
Extraordinary claims need more than a four-page note. This one has several things going for it.
First, the person. AlpΓΆge is a serious number theorist, not a crank with a website. Second, the track record on the same territory. In August 2026, Anthropic reported that an experimental version of Claude had improved a long-standing result about the zeta function, raising the proven lower bound on the proportion of zeros lying on the critical line from 41.6% to 67.25%. That was not the Riemann Hypothesis being "two-thirds solved" β the full hypothesis concerns every relevant zero, and closing the remaining gap requires a proof that covers them all β but mathematicians examined the earlier result, and Anthropic released both a paper and a formalization. And in July 2026, AlpΓΆge announced a Claude-assisted counterexample to the Jacobian conjecture, another long-standing problem. This claim arrives against a background of substantive, checked work on the same mathematical ground.
Third, the early reactions from people qualified to judge. James Maynard, the Regius Professor of Mathematics at Oxford and a 2022 Fields Medal winner, told IBM Think: "If this fully checks out, it is clearly a historic milestone in mathematics, establishing one of the great mysteries of the subject." He added: "I suspect that the proof will turn out to be correct, although even checking the Lean verification will take some time."
Lamzouri, who developed a shorter, simpler proof of the earlier 67.25% result without AI, told IBM Think he was astonished: "I don't have words to describe what I am feeling right now. I am just in awe. If this claim happens to be true, it will be a huge achievement." And Brian Conrey, a leading specialist on the Riemann Hypothesis, told AFP: "It's a monumental moment in human history. It's really the biggest thing I think that's ever happened in mathematics." He was careful to add that he had not yet seen a manuscript, only AlpΓΆge's remarks, and had asked for a copy.
That mix β genuine excitement from the top of the field, paired with open admissions that nobody has checked anything yet β is exactly the right posture. It is also the posture this article is taking.
The Stanford twist
Hours after AlpΓΆge's announcement, there was a second claim. Stanford mathematics professor Ravi Vakil said a postdoctoral researcher in the same department had independently submitted a proof of the hypothesis. Neither Vakil nor the researcher immediately responded to requests for comment, so almost nothing is known about that second claim yet β not the method, not the format, not whether it overlaps with AlpΓΆge's. But it means the verification story now has two threads to follow, and possibly a priority question if both survive scrutiny.
The IPO timing
You can't write about this announcement without writing about its timing, because AlpΓΆge himself raised it. The announcement came the day after Anthropic's record stock-market debut β an IPO with roughly $93.5 billion in gross proceeds, shares closing their first day at $97.17, about 17% above the IPO price. AlpΓΆge said he had consulted Anthropic chief executive Dario Amodei before announcing, and acknowledged that the timing could be viewed cynically, without elaborating. Anthropic declined to comment.
The honest way to handle this: the timing is genuinely awkward, and noticing that doesn't make you a cynic. It also doesn't make the proof wrong. A Lean formalization either compiles and states the theorem or it doesn't, and no IPO changes that. The right response is to hold both thoughts at once β the timing invites questions, and the mathematics will answer them on its own schedule.
There is also recent history worth knowing. Two months ago, OpenAI said an AI model had solved another Millennium Prize Problem, on the equations that govern how fluids move. Mathematicians say that proof appears correct but has yet to be fully vetted. AlpΓΆge, who had worked on related fluid problems with NYU mathematician Tristan Buckmaster, was at the center of a dispute over credit for that work; OpenAI denied wrongdoing. One observer told AFP he was glad that "this time there might not be any of the same corporate controversy," and praised AlpΓΆge for being open that he did not yet fully understand the proof β the back-and-forth, he said, showed AI working "more like a tool" and helped dispel "this idea of the machine god."
What it would mean if it holds
If verified, this solves one of the seven Millennium Prize Problems and earns the $1 million Clay Mathematics Institute prize. But the prize itself is slow: the Clay Institute's rules require publication in a qualifying outlet, at least two years after publication, and general acceptance by the global mathematics community before it will consider a proposed solution. Nobody is collecting a cheque this year.
The mathematical consequences would be far larger than the money. A proof of the Riemann Hypothesis would tighten our understanding of how primes are distributed and settle thousands of results that currently rest on the assumption that it's true. The Generalized version would extend that to a whole family of related functions.
One thing it would not do, at least not directly: break your encryption. The Riemann Hypothesis is about the distribution of primes, not about factoring large numbers, and the leap from "we understand primes better" to "RSA is broken" is not one any serious mathematician is making. If you see that claim in the next few days, treat it as a sign the writer hasn't done the reading.
The bigger question, for people who build software rather than prove theorems, is what this says about AI as a research partner. AlpΓΆge describes a collaboration in which he directed the effort toward a promising approach and Claude supplied major advances as the argument developed β including, by his account, pulling ahead near the finish. His sketch draws on a technique used in another mathematical setting where an analogue of the Riemann Hypothesis is already known; the difficulty is carrying the useful idea across to ordinary integers. According to the note, Claude helped find a workaround involving a long chain of transformations called theta lifts. Specialists will have to assess that explanation. For everyone else, the important detail is that AlpΓΆge attributes mathematical ideas to Claude, including a central step in the proposed argument.
If that holds up, it is a specific, inspectable example of a human mathematician and a model reaching a result that generations of researchers couldn't β and a case study in how the human direction and the model's contributions fit together. That is worth studying even if you never touch number theory.
What to watch next
- Does it compile? The first independent check is mechanical: do the 14 archive parts build, and does the final theorem actually state the Riemann Hypothesis? Watch for reports from Lean specialists in the coming days.
- What does it depend on? The second check is about the foundations: which axioms and assumptions the formalization relies on, and whether any of them smuggle in the result.
- The Stanford claim. Almost nothing is known about it yet. If details emerge, the two claims need to be compared β and the priority question matters.
- A readable write-up. AlpΓΆge intends to understand the argument more fully and produce a proper written account. That is the document the mathematical community actually needs before it can accept the result in the human sense, and it will take time.
- The Clay clock. Even in the best case, the prize process takes years: qualifying publication, two years of scrutiny, general acceptance. Treat any headline about the million dollars as premature until that process starts.
Our take
This is the most exciting unproven claim in mathematics in living memory, and also the most honestly uncertain one. The ingredients for taking it seriously are all there: a credible mathematician, a checkable formalization, a track record on the same problem, and early optimism from Fields Medalists. The ingredients for proof are not there yet, and won't be for a while β possibly months, possibly years, possibly never, if a flaw turns up in the formalization.
The correct position right now is excitement plus scepticism, held at the same time. Don't dismiss it because an AI was involved, and don't celebrate it because a headline told you to. The mathematics will decide, and the Lean files are public β which means the deciding is already underway, in the open, by people who know what they're looking at.
The next report worth reading isn't another announcement. It's the first one from a researcher who has checked the files and can begin explaining what Claude and AlpΓΆge actually found.
Sources: IBM Think, "Anthropic mathematician claims AI helped solve a 167-year-old math mystery," November 20, 2026 (https://www.ibm.com/think/news/anthropic-claude-riemann-hypothesis-math-proof); eWeek, "Claude and a 167-Year-Old Math Mystery: The Riemann Hypothesis Proof Claim," November 20, 2026 (https://www.eweek.com/news/claude-riemann-hypothesis-proof-claim/); The Daily Star / AFP, "Claude solved math's greatest riddle, says Anthropic researcher," November 21, 2026 (https://www.thedailystar.net/news/world/news/claude-solved-maths-greatest-riddle-says-anthropic-researcher-4329281); The Wall Street Journal, "Anthropic Mathematician Says Claude Has Solved the Most Famous Problem in Math" (https://www.wsj.com/tech/ai/anthropic-mathematician-says-claude-has-solved-the-most-famous-problem-in-math-9361920b); Levent AlpΓΆge's announcement note (https://alpo.ge). The figures for the Lean formalization's size (around 17 million lines across 14 archive parts) and the full written argument (more than 250,000 pages) are as reported in the announcement coverage; the IPO figures (about $93.5 billion gross proceeds, first-day close $97.17, about 17% above the IPO price) are as reported. The claim has not been independently verified and has not been peer-reviewed; where this article says a fact is unknown, that is because it has not been published, not because it was missed. No hands-on testing was performed for this article. This breaking article is not the daily buyer-intent post, and it contains no affiliate links. AI disclosure: this article was drafted with AI assistance and checked against the IBM, eWeek and Daily Star reports.
Also on ToolsDig: Claude May Have Just Proved the Riemann Hypothesis. Here's Why You Shouldn't Celebrate Yet
Originally published by Dev.to AI. Aggregated on AIWithGhost for educational purposes β full credit and traffic to the original publisher.