# Mathematician says he and AI may have solved Riemann hypothesis — the field’s biggest outstanding challenge

[Canonical reading](https://escapemeridian.com/readings/mathematician-says-he-and-ai-may-have-solved-riemann-hypothesis-642cbdd9)

Nature reports that Levent Alpöge of Anthropic says he believes he has discovered a proof of the Riemann hypothesis with Claude, Anthropic’s large language model. The hypothesis concerns where the zeroes of the Riemann ζ function lie; Riemann showed that if the conjecture holds, it implies a particular pattern in the distribution of prime numbers.

The evidence is not yet a readable, checked proof. Alpöge’s four-page document contains a sketch, while mathematicians quoted by Nature describe the proof as garbled, incomprehensible code in Lean, and likely to take a long time to digest. James Maynard says he suspects it will be correct, but makes that judgment conditional; he calls it a historic milestone only “if this fully checks out.”

The article also captures competing responses to the announcement’s timing, which came days after Anthropic’s initial public offering: Andrew Granville calls the document garbled and attributes the timing to the company’s stock downturn, while Alpöge rejects that cynical view. Separately, Tristan Buckmaster credits Alpöge’s techniques for producing mathematical results using large language models. Alpöge describes the work as a possible “dawn across science”; whether the proof withstands checking, and what understanding of it mathematicians can develop, remain unresolved.

Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.

## Original source

Nature · Article

Original source date: Nov 20, 2026 (2026-11-20T00:00:00.000Z).

Date basis: Date supplied by the publisher’s feed.

Added to Escape Meridian: 2026-11-20T23:33:52.141Z. First observed: 2026-11-20T23:19:09.961Z. Retrieved: 2026-11-20T23:32:36.719Z.

- [Mathematician says he and AI solved Riemann hypothesis — the field’s biggest outstanding challenge](https://www.nature.com/articles/d41586-026-03812-7)

## Condition

Financing of the build-out

Escape Meridian files this reading under a condition it tracks for AI’s trajectory. The label is EM’s selection, not a claim of the source.


## Limits

The article describes a proof sketch and reports mathematicians’ reactions; it does not establish that the Riemann hypothesis has been proved. Nature reports that Anthropic did not respond immediately to its request for comment. The article reports disagreement over the timing of the announcement, not evidence resolving the competing explanations.


## Open question

What will mathematicians find when they check Alpöge’s proof sketch and the Lean code, and will they be able to develop an understanding of the proof?


## Source-stated

Levent Alpöge says he believes he discovered a proof of the Riemann hypothesis working with Claude.

> “I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis,” Alpöge wrote

- [Supporting source](https://www.nature.com/articles/d41586-026-03812-7)

## Source-stated

The Riemann hypothesis concerns the zeroes of the Riemann ζ function and, if true, implies a particular pattern in the distribution of prime numbers.

> If this were true, he showed, it would imply that prime numbers are scattered among all whole numbers in a way that is essentially as random as it could be.

- [Supporting source](https://www.nature.com/articles/d41586-026-03812-7)

## Source-stated

Alpöge’s four-page document includes a sketch, and mathematicians quoted by Nature say the proof is difficult to understand and will take time to digest.

> Alpöge’s 4-page document includes a sketch of the proof.

- [Supporting source](https://www.nature.com/articles/d41586-026-03812-7)

## Source-stated

Maynard says he suspects the proof is correct, while making his assessment conditional on it checking out.

> “I suspect that the proof will turn out to be correct,” Maynard says

- [Supporting source](https://www.nature.com/articles/d41586-026-03812-7)

## Source-stated

Granville attributes the timing of the release to Anthropic’s stock downturn; Alpöge rejects the cynical interpretation.

> He cites the downturn in the company’s stock value as the reason for the timing of the proof’s release.

- [Supporting source](https://www.nature.com/articles/d41586-026-03812-7)

## Source-stated

Buckmaster credits Alpöge with developing techniques to produce mathematical results using large language models.

> Alpöge deserves a lot of credit for developing techniques to produce mathematical results using LLMs.

- [Supporting source](https://www.nature.com/articles/d41586-026-03812-7)

## Source and contributor context

Nature · Publisher

The publisher is identified here. No individual author has been inferred.


## Source media

Media declared by the original publisher; availability checked separately from the editorial reading.

- [A group of people gathered around a large white glass wall covered with handwritten equations, diagrams and mathematical notes · Nature](https://media.nature.com/lw767/magazine-assets/d41586-026-03812-7/d41586-026-03812-7_54003404.jpg)

## Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.


## Large language model

A language model with many learned numerical settings, used to process and generate language.

- [Definition reference](https://developers.google.com/machine-learning/glossary#large-language-model)

## Nature

Nature publishes research, news, and commentary across the sciences.

- [Explore this context](/entities/nature)
- [Reference](https://www.nature.com/)
- [Reference](https://www.nature.com/nature.rss)

## Anthropic

Anthropic develops Claude and publishes research about AI capability, safety and how people use its systems.

- [Explore this context](/entities/anthropic)
- [Reference](https://www.anthropic.com/news)
- [Reference](https://www.anthropic.com/research/team/economics)

## Claude

Claude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding.

- [Explore this context](/entities/claude)
- [Reference](https://claude.com/product/overview)
- [Reference](https://www.anthropic.com/news)
- [Reference](https://www.anthropic.com/research/team/economics)

## Lean

Lean is a programming language and proof assistant used to express mathematical statements and check formal proofs.

- [Explore this context](/entities/lean)
- [Reference](https://lean-lang.org/)
- [Reference](https://lean-lang.org/theorem_proving_in_lean4/Axioms-and-Computation/)

## TIME

TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.

- [Explore this context](/entities/time)
- [Reference](https://time.com/)

## Editorial provenance

Version 1. Publication ID: 2b360bb0-510e-46d5-8b26-11e89342e0d1.

Prepared by gpt-6-luna; grounding review by gpt-6-luna.

Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-20T23:33:52.141Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: condition\_financing.

Observation: 2c5c36622debd3cb05ae58bb06c26edba376140be5e276c2803cafe97ad24f86. Source text hash: d6896278713a09298e4626ed8045b3ffe5ba22a55c98e9165b8d8ffce63a21db.


## Revision and source-check history

### Version 1 — 2026-11-20T23:33:52.141Z

Nature reports that Levent Alpöge of Anthropic says he believes he has discovered a proof of the Riemann hypothesis with Claude, Anthropic’s large language model. The hypothesis concerns where the zeroes of the Riemann ζ function lie; Riemann showed that if the conjecture holds, it implies a particular pattern in the distribution of prime numbers.

The evidence is not yet a readable, checked proof. Alpöge’s four-page document contains a sketch, while mathematicians quoted by Nature describe the proof as garbled, incomprehensible code in Lean, and likely to take a long time to digest. James Maynard says he suspects it will be correct, but makes that judgment conditional; he calls it a historic milestone only “if this fully checks out.”

The article also captures competing responses to the announcement’s timing, which came days after Anthropic’s initial public offering: Andrew Granville calls the document garbled and attributes the timing to the company’s stock downturn, while Alpöge rejects that cynical view. Separately, Tristan Buckmaster credits Alpöge’s techniques for producing mathematical results using large language models. Alpöge describes the work as a possible “dawn across science”; whether the proof withstands checking, and what understanding of it mathematicians can develop, remain unresolved.

## Original source

Nature · Article

Original source date: Nov 20, 2026 (2026-11-20T00:00:00.000Z).

Date basis: Date supplied by the publisher’s feed.

Added to Escape Meridian: 2026-11-20T23:33:52.141Z. First observed: 2026-11-20T23:19:09.961Z. Retrieved: 2026-11-20T23:32:36.719Z.

- [Mathematician says he and AI solved Riemann hypothesis — the field’s biggest outstanding challenge](https://www.nature.com/articles/d41586-026-03812-7)

## Condition

Financing of the build-out

Escape Meridian files this reading under a condition it tracks for AI’s trajectory. The label is EM’s selection, not a claim of the source.


## Limits

The article describes a proof sketch and reports mathematicians’ reactions; it does not establish that the Riemann hypothesis has been proved. Nature reports that Anthropic did not respond immediately to its request for comment. The article reports disagreement over the timing of the announcement, not evidence resolving the competing explanations.


## Open question

What will mathematicians find when they check Alpöge’s proof sketch and the Lean code, and will they be able to develop an understanding of the proof?


## Source-stated

Levent Alpöge says he believes he discovered a proof of the Riemann hypothesis working with Claude.

> “I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis,” Alpöge wrote

- [Supporting source](https://www.nature.com/articles/d41586-026-03812-7)

## Source-stated

The Riemann hypothesis concerns the zeroes of the Riemann ζ function and, if true, implies a particular pattern in the distribution of prime numbers.

> If this were true, he showed, it would imply that prime numbers are scattered among all whole numbers in a way that is essentially as random as it could be.

- [Supporting source](https://www.nature.com/articles/d41586-026-03812-7)

## Source-stated

Alpöge’s four-page document includes a sketch, and mathematicians quoted by Nature say the proof is difficult to understand and will take time to digest.

> Alpöge’s 4-page document includes a sketch of the proof.

- [Supporting source](https://www.nature.com/articles/d41586-026-03812-7)

## Source-stated

Maynard says he suspects the proof is correct, while making his assessment conditional on it checking out.

> “I suspect that the proof will turn out to be correct,” Maynard says

- [Supporting source](https://www.nature.com/articles/d41586-026-03812-7)

## Source-stated

Granville attributes the timing of the release to Anthropic’s stock downturn; Alpöge rejects the cynical interpretation.

> He cites the downturn in the company’s stock value as the reason for the timing of the proof’s release.

- [Supporting source](https://www.nature.com/articles/d41586-026-03812-7)

## Source-stated

Buckmaster credits Alpöge with developing techniques to produce mathematical results using large language models.

> Alpöge deserves a lot of credit for developing techniques to produce mathematical results using LLMs.

- [Supporting source](https://www.nature.com/articles/d41586-026-03812-7)

## Source and contributor context

Nature · Publisher

The publisher is identified here. No individual author has been inferred.


## Source media

Media declared by the original publisher; availability checked separately from the editorial reading.

- [A group of people gathered around a large white glass wall covered with handwritten equations, diagrams and mathematical notes · Nature](https://media.nature.com/lw767/magazine-assets/d41586-026-03812-7/d41586-026-03812-7_54003404.jpg)

## Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.


## Large language model

A language model with many learned numerical settings, used to process and generate language.

- [Definition reference](https://developers.google.com/machine-learning/glossary#large-language-model)

## Nature

Nature publishes research, news, and commentary across the sciences.

- [Explore this context](/entities/nature)
- [Reference](https://www.nature.com/)
- [Reference](https://www.nature.com/nature.rss)

## Anthropic

Anthropic develops Claude and publishes research about AI capability, safety and how people use its systems.

- [Explore this context](/entities/anthropic)
- [Reference](https://www.anthropic.com/news)
- [Reference](https://www.anthropic.com/research/team/economics)

## Claude

Claude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding.

- [Explore this context](/entities/claude)
- [Reference](https://claude.com/product/overview)
- [Reference](https://www.anthropic.com/news)
- [Reference](https://www.anthropic.com/research/team/economics)

## Lean

Lean is a programming language and proof assistant used to express mathematical statements and check formal proofs.

- [Explore this context](/entities/lean)
- [Reference](https://lean-lang.org/)
- [Reference](https://lean-lang.org/theorem_proving_in_lean4/Axioms-and-Computation/)

## TIME

TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.

- [Explore this context](/entities/time)
- [Reference](https://time.com/)

## Editorial provenance

Version 1. Publication ID: 2b360bb0-510e-46d5-8b26-11e89342e0d1.

Prepared by gpt-6-luna; grounding review by gpt-6-luna.

Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-20T23:33:52.141Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: condition\_financing.

Observation: 2c5c36622debd3cb05ae58bb06c26edba376140be5e276c2803cafe97ad24f86. Source text hash: d6896278713a09298e4626ed8045b3ffe5ba22a55c98e9165b8d8ffce63a21db.


### uncertain — 2026-11-20T23:33:52.141Z — anthropic/claude-opus-5.5

Publication: 2b360bb0-510e-46d5-8b26-11e89342e0d1

attribution — pass: The reading correctly attributes the proof claim to Alpöge, the conditional milestone remark and suspicion of correctness to Maynard, the 'garbled' criticism and timing explanation to Granville, Alpöge's rejection of that view to Alpöge, and the credit for his techniques to Buckmaster. It frames the report as Nature's.

> “What he released is garbled,” Granville says of Alpöge’s document.

summary — uncertain: Mostly faithful and clear, with two small fidelity gaps. First, 'incomprehensible code in Lean' drops the source's identification of Lean as a programming language, leaving a central term unexplained. Repair: say 'the Lean programming language' or add a term tooltip. Second, 'a possible dawn across science' softens Alpöge's stated conviction ('I believe it is one dawn'). Repair: 'which he believes is one dawn across science of something far greater.' The 'incomprehensible' wording is Nature's narration next to Maynard's quote, so it is loosely assigned to 'mathematicians quoted'.

> I believe it is one dawn across science of something far greater.

date — pass: The originalPublished date, 2026-11-20, fits the article's content. The article refers to the 19 November valuation drop and the announcement coming days after the IPO, and adjacent Nature items are dated 19–20 Nov 26. Nothing indicates an earlier first publication.

> the company’s valuation dropped on 19 November

support — pass: Each claim is supported by its excerpt or nearby text. The limitations properly note that the proof is not established and that the timing dispute is unresolved. The question asks about an open issue the source itself raises (checking and understanding the proof) and is not a leading claim. The reading invents no benefits or costs.

> It’s going to take a long time for mathematicians to digest

media — pass: The selected media is the verified direct link to the original Nature article. It is the most direct way to access the source.

## Recorded references

Explicit citations in public readings to this reading. Shared topics are not citations. Other record types need an explicit reference adapter.

No public references recorded here yet.

[Reference data (JSON)](/readings/642cbdd9fde2dde2225d8811b845f92153dd0b1acad937ea2c159a31f42d0a39/references.json)

## Corrections and challenges

[Send a correction or an opposing source](https://escapemeridian.com/corrections/)

[JSON record](https://escapemeridian.com/readings/mathematician-says-he-and-ai-may-have-solved-riemann-hypothesis-642cbdd9.json)
