ResearchPublished All news from that day
Ongoing case: AI-written mathematics and how it is checked 6 storiesAnthropic's Levent Alpöge believes he and Claude found a Riemann hypothesis proof
Alpöge writes that he believes the proof also covers the generalized hypothesis, and says he posted only Lean files, not a written paper.
Check our sources · 11 facts from 3 sourcesLevent Alpöge, a mathematician at Anthropic, writes: "I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis." The claim comes from a short PDF on his own website, and it is stated as his belief.
He says the method extends to the Generalized Riemann Hypothesis, and that he expects, "with more time," the Extended Riemann Hypothesis too. He adds that he and Claude "have also formalized the arguments without assumptions."
He chose not to post a written argument. Alpöge writes that it would be "obviously inappropriate to attempt to post some massive unreadable travesty," so he decided "to only present Lean", a proof assistant that checks every step of a proof by computer. A footnote points to the files, split into 14 parts on his site.
He credits Claude with most of the final step: "this time Claude far outpaced me near the finish line." The PDF names only "Claude", with no model version.
This is not Anthropic's first Riemann result. Its own research page of 10 August says an unreleased research version of Claude raised a known lower bound on the share of zeta zeros on the critical line from 41.6% to 67.2%, and that Anthropic did not expect those techniques to prove the hypothesis. Alpöge is one of the two Anthropic mathematicians named there as having examined that work.
Underlined sentences link to the facts we checked. Dotted words open a short explainer.
Fact 1 of 11 · from Levent Alpöge, personal site, 20 Nov 2026
I believe I have, working with Claude, discovered a proof1 of the Riemann Hypothesis.
Fact 4 of 11 · from Levent Alpöge, personal site, 20 Nov 2026
It being obviously inappropriate to attempt to post some massive unreadable travesty of "slop", I decided to only present Lean.
Your reaction
We count reactions per story and day, never who reacted. The counts help us choose what goes in the monthly issue. If you are signed in, your own page shows yours too.
We count reactions per story and day, never who reacted. They help choose what goes in the monthly issue.
Check our sources
Every sentence above is checked against these 3 sources.
1 remark (Levent Alpöge)
Open the source archived copy-
Levent Alpöge writes in his PDF "remark": "I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis."
citeI believe I have, working with Claude, discovered a proof1 of the Riemann Hypothesis.
-
Alpöge writes that the method "easily extends much further, resolving the so-called Generalized Riemann Hypothesis, and I expect, with more time, the Extended Riemann Hypothesis too."
citethe method easily extends much further, resolving the so-called Generalized Riemann Hypothesis, and I expect, with more time, the Extended Riemann Hypothesis too.
-
Alpöge writes: "Claude and I have also formalized the arguments without assumptions."
citeClaude and I have also formalized the arguments without assumptions.
-
Alpöge writes that it would be "obviously inappropriate to attempt to post some massive unreadable travesty of \"slop\"" and that he therefore decided "to only present Lean".
citeIt being obviously inappropriate to attempt to post some massive unreadable travesty of "slop", I decided to only present Lean.
-
A footnote in Alpöge's PDF gives the proof's files as alpo.ge/lean.tar.gz.part01 through alpo.ge/lean.tar.gz.part14.
cite1https://alpo.ge/lean.tar.gz.part01, . . . , https://alpo.ge/lean.tar.gz.part14.
-
Alpöge writes of the final stage of the work: "this time Claude far outpaced me near the finish line."
citebut this time Claude far outpaced me near the finish line.
-
Alpöge's PDF names the AI only as "Claude" and gives no model version.
citeClaude and I have had a great time together
2 riemann zeta
Open the source archived copy-
Anthropic's research page of 10 August 2026 names Levent Alpöge as one of two Anthropic mathematicians who examined Claude's earlier work on the Riemann zeta function.
citeLevent Alpöge and Ralph Furman, two of Anthropic’s own mathematicians, examined Claude’s work
-
Anthropic's research page of 10 August 2026 says "an unreleased research version of Claude" improved a longstanding lower bound for the fraction of zeros of the Riemann zeta function that satisfy the Riemann hypothesis, raising it from 41.6% to 67.2%.
Toned down to what the source says
citeAn unreleased research version of Claude has improved on a longstanding lower bound for the fraction of zeros of the Riemann zeta function that satisfy the Riemann hypothesis. ... it has increased this bound from 41.6% to 67.2%.
-
Anthropic's research page of 10 August 2026 says it does not expect the techniques used to lead to a proof of the Riemann hypothesis.
citeWe don’t expect that the techniques Claude used will lead to proving the Riemann hypothesis.
3 Lean programming language and theorem prover
Open the source archived copy-
Lean describes itself as an open-source programming language and proof assistant, and its home page says: "Define it, prove it, and let Lean check every step."
cite
Every sentence above is checked against these 3 sources. Each fact has its own address you can cite.
Source 1 · Levent Alpöge, personal site · 20 Nov 2026
remark (Levent Alpöge)
Levent Alpöge writes in his PDF "remark": "I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis."
I believe I have, working with Claude, discovered a proof1 of the Riemann Hypothesis.
Alpöge writes that the method "easily extends much further, resolving the so-called Generalized Riemann Hypothesis, and I expect, with more time, the Extended Riemann Hypothesis too."
the method easily extends much further, resolving the so-called Generalized Riemann Hypothesis, and I expect, with more time, the Extended Riemann Hypothesis too.
Alpöge writes: "Claude and I have also formalized the arguments without assumptions."
Claude and I have also formalized the arguments without assumptions.
Show all 7 factsShow fewer facts
Alpöge writes that it would be "obviously inappropriate to attempt to post some massive unreadable travesty of \"slop\"" and that he therefore decided "to only present Lean".
It being obviously inappropriate to attempt to post some massive unreadable travesty of "slop", I decided to only present Lean.
A footnote in Alpöge's PDF gives the proof's files as alpo.ge/lean.tar.gz.part01 through alpo.ge/lean.tar.gz.part14.
1https://alpo.ge/lean.tar.gz.part01, . . . , https://alpo.ge/lean.tar.gz.part14.
Alpöge writes of the final stage of the work: "this time Claude far outpaced me near the finish line."
but this time Claude far outpaced me near the finish line.
Alpöge's PDF names the AI only as "Claude" and gives no model version.
Claude and I have had a great time together
Source 2 · Anthropic · 10 Aug 2026
riemann zeta
Anthropic's research page of 10 August 2026 names Levent Alpöge as one of two Anthropic mathematicians who examined Claude's earlier work on the Riemann zeta function.
Levent Alpöge and Ralph Furman, two of Anthropic’s own mathematicians, examined Claude’s work
Anthropic's research page of 10 August 2026 says "an unreleased research version of Claude" improved a longstanding lower bound for the fraction of zeros of the Riemann zeta function that satisfy the Riemann hypothesis, raising it from 41.6% to 67.2%.
Toned down to what the source saysAn unreleased research version of Claude has improved on a longstanding lower bound for the fraction of zeros of the Riemann zeta function that satisfy the Riemann hypothesis. ... it has increased this bound from 41.6% to 67.2%.
Anthropic's research page of 10 August 2026 says it does not expect the techniques used to lead to a proof of the Riemann hypothesis.
We don’t expect that the techniques Claude used will lead to proving the Riemann hypothesis.
Source 3 · Lean FRO
Lean programming language and theorem prover
Lean describes itself as an open-source programming language and proof assistant, and its home page says: "Define it, prove it, and let Lean check every step."
Topics
The morning email
On the mornings we publish: the three top stories and up to four short ones. Free.
We email you a link to confirm. An issue may include one sponsor, always labelled Sponsored · Advertisement. Our emails count opens and clicks, not who made them. Unsubscribe in one click. What we keep
