Anthropic Researcher Claims to Have Proven the Riemann Hypothesis with AI. Is the 167-Year-Old Math Problem Finally Solved?
The Riemann hypothesis, which has bedeviled the mathematics world for 167 years, may be on the verge of its first major breakthrough — but whether an AI-generated proof can withstand scrutiny from the mathematics community remains the biggest open question.
On November 20, Anthropic mathematician Levent Alpöge issued a statement claiming that, working with the AI model Claude, he had found a proof of the Riemann hypothesis and the generalized Riemann hypothesis (GRH), and simultaneously released more than 17 million lines of Lean formalization code. Because the announcement coincided with the day after Anthropic went public, the mathematical breakthrough also sparked outside discussion about the timing of the release and its commercial motives.
Alpöge wrote the proof in the Lean language and released it in 14 compressed files. At the same time, the text version of the proof he generated with Claude exceeds 250,000 pages, but it was not made public because it is too long and difficult to read.
At present, no mathematician has formally confirmed that the proof holds. It still needs to be verified externally whether the theorem ultimately proven by the Lean code actually corresponds to the Riemann hypothesis itself, and whether there are logical gaps anywhere in the proof.
Classic portrait of Riemann (1850s)
The Proof Method: Claude's Key Breakthrough "at the Finish Line"
In the statement, Alpöge described the core of the collaboration: he pointed Claude toward the key research direction, and Claude then proposed what he considers a highly groundbreaking key step — a "theta lifting" chain whose length is roughly equal to the square root of the sum of the bounded primes. Each lifting corresponds to the "multiplicity" on which classical function field arguments rely, and that multiplicity is extremely hard to find over the integers.
"Approaching the finish line, Claude far surpassed me. No, Claude was simply flying, dancing, and shining near the finish line," Alpöge wrote.
He also attached a research timeline said to have been compiled by Claude: the work began on October 6 with the "Stepanov determinant," and by November 16 only one entry remained — "GRH."
MIT mathematician Andrew Sutherland, after independently compiling the relevant Lean code, said: "This is indeed a historic moment. The world has already changed... in my view, this is a victory for humanity, a day worth celebrating."
Reaction from the Math World: Cautious Optimism, Awaiting Independent Verification
Many top mathematicians and AI researchers expressed interest in the development, but remained cautious about whether the proof truly holds.
Terence Tao, a 2006 Fields Medal winner, wrote on his personal blog that a preliminary analysis shows the proof appears to draw on modern results from the Langlands program and arithmetic geometry in combination, and verification is expected to take some time.
Mathematician Elliot Glazer's preliminary check found that the mathematical proposition formalized in the Lean file is indeed the generalized Riemann hypothesis, and he estimated the probability that Lean itself contains a flaw to be "far below 1%." Alex Kontorovich, chair of the mathematics department at Rutgers University, also said that as more details emerge, he is more confident in the likelihood that the proof succeeds.
A mathematics and AI researcher at Princeton University emphasized that human involvement remains the best way to use large language models to solve mathematical problems. He also noted that in this research, Alpöge should receive credit equal to Claude's.
A 250,000-Page Proof: The Conflict Between AI Breakthroughs and Human Verification
The sheer length of the proof has become another focus of attention in the mathematics world.
University of Manchester mathematician Thomas Bloom said on X that he did not know whether he was more stunned by AI claiming to have proven the Riemann hypothesis, or by AI claiming to have generated a logically coherent proof that exceeds 250,000 pages once converted into human-readable form.
Pierre Richemond, head of Cohere's multimodal business and a former Google DeepMind researcher, argued that the Riemann hypothesis may already have been solved, but it still awaits Lean verification. The development raises more questions: AI-generated research output already carries serious mathematical value, yet it has not been fully verified by the human mathematics community.
Timing Controversy: Announced the Day After the IPO, Alpöge Responds Proactively
The timing of the announcement also drew outside attention. Anthropic went public on November 19, and Alpöge disclosed this major mathematical advance the next day, inevitably prompting speculation in the market about motives of commercial promotion.
Alpöge addressed the controversy directly in his statement, acknowledging that "the exact timing is very bad." He said the key work was completed from late at night on November 15 to the early hours of November 16 local time, after which he asked Anthropic CEO Dario Amodei whether he could make it public; "eventually we figured out that there were indeed restrictions that had to be followed."
He denied negative outside speculation about the release timing, saying such an "obviously cynical view" is wrong, and said he "simply could not wait any longer," and was willing to take full responsibility for it.
It is worth noting that Anthropic had previously made incremental progress in AI mathematics research. In August this year, the company published Claude's research results on the distribution of zeros of the Riemann zeta function, raising the lower bound on the proportion of zeros on the critical line from 41.6% to 67.2%. At the time, however, Anthropic explicitly stated that it did not expect the techniques Claude used to be able to ultimately prove the Riemann hypothesis.
Prize and Recognition: Still Years Away from the $1 Million Award
Even if the proof ultimately holds, Alpöge remains a considerable distance from the $1 million prize established by the Clay Mathematics Institute.
Under the institute's rules, a solution to a Millennium Prize Problem must be formally published, undergo at least two years of public scrutiny, and gain broad recognition from the international mathematics community before a prize can be applied for.
Alpöge himself is optimistic about what the proof will look like once it is ultimately accepted. He believes the proof will very likely be greatly simplified, and that this may be achievable soon. He described the development as "a dawn of something greater in the field of science."
Capital markets have not yet formed a clear judgment on the news either. Relevant odds on prediction platforms still reflect a divided market. Meanwhile, an old 2022 post on X by OpenAI CEO Sam Altman, half-jokingly asking "when will the Riemann hypothesis be done," has been dug up by netizens, with someone joking in the comments: "Anthropic beat you to it, bro."
Join the official Coincamps community:
Telegram: https://t.me/coin_camps
Recommended
FNB Alternatives (2026): 5 Banks That Charge Less When a Debit Order Bounces
Nov 22, 01:35
22K vs 24K Gold (2026): Purity, Price and Resale Compared
Nov 21, 23:35
The Easy-Money Window Has Closed: Crypto's Four Liquidity Harvesting Machines
Nov 21, 22:40Large USDT Savings From Before Marriage: Must It Be Split 50/50 in a Divorce?
Nov 21, 20:55CZ's Japan Interview: From Binance to Giggle Academy, on Entrepreneurship, AI and Young People's Career Choices
Nov 21, 18:40Road Tax 2027: New UK Car Tax Rates From April and What You'll Pay
Nov 21, 17:30