Anthropic mathematician says Claude helped prove the Riemann hypothesis
The most famous open problem in math is 167 years old. An Anthropic employee says he cracked it with Claude, but checking has only begun.
Event: November 20, 2026
Primary source: November 20, 2026
Published here: November 21, 2026 at 12:57 AM UTC
Read: 2 min
What happened
On November 20 Anthropic mathematician Levent Alpöge posted a short note on his website. He says that, working with Claude, he found a proof of the Riemann hypothesis and of its generalized version too. He released the proof not as a paper but as code in Lean, software that checks math step by step.
Why it matters to you
If the proof holds, it is the biggest AI result in science so far, made by the model of a company that just went public. Claims like this matter for how AI companies are valued. But famous proofs of this hypothesis have appeared before, and all of them failed.
What to check and watch
What independent mathematicians say: whether the Lean statement matches the real hypothesis and whether there are errors. The claim is not accepted yet.
Where it’s happening
LIVEDrag to spin. Green dots: the exchange is trading right now (regular hours, holidays not counted). Shadow: where it is night now. Borders: Natural Earth, quakes: USGS. Layer idea: God’s Eye View (MIT).
Flights: what adsb.lol volunteer ADS-B receivers see right now (ODbL), refreshed every minute. Few receivers over oceans and remote land; military and private jets are often hidden.
What the problem is. Bernhard Riemann posed the hypothesis in 1859. It is about how prime numbers, the ones divisible only by 1 and themselves, are spread out. The Clay Mathematics Institute lists it among its Millennium Prize Problems with a $1 million award. Many thought a solution was still very far away.
What exactly is claimed. Alpöge writes that the method also settles the generalized Riemann hypothesis and that the proof is fully formalized in Lean. The archive comes in 14 parts. He admits the argument is very long right now and expects it to be simplified a lot. There is no regular paper yet.
First checks. MIT mathematician Andrew Sutherland independently compiled Alpöge's Lean code, Scientific American reports, and called it a historic day. But code that compiles does not settle the problem: experts must confirm that the Lean statement really is the Riemann hypothesis and work through the proof. To win the Clay prize, a solution must appear in a qualifying outlet, survive at least two years of scrutiny and win general acceptance.
Why the timing raises questions. Alpöge writes that he messaged Anthropic CEO Dario Amodei late on Sunday, November 15, and that there was a constraint to respect. The note came out the day after Anthropic's first day of trading on Nasdaq. He calls the timing "horrible" and says the cynical reading is false.
What it means for you. The hypothesis has no direct everyday use: banking and the internet will not change because of it. For the AI market it is a signal of what models can do. Do not buy or sell stocks on one claim; wait for the checks.
Confirmed facts
- On November 20 Levent Alpöge, a mathematician at Anthropic, posted a note saying that, working with Claude, he found a proof of the Riemann hypothesis.[1][2]
- He says the method also resolves the generalized Riemann hypothesis.[1]
- The proof is formalized in Lean and posted as an archive in 14 parts.[1][3]
- Alpöge writes that the argument is very long for now and he expects it to be simplified a lot.[1]
- Bernhard Riemann posed the hypothesis in 1859; the Clay Mathematics Institute offers $1 million for a solution.[2]
- MIT mathematician Andrew Sutherland independently compiled Alpöge's Lean code.[2]
- For the Clay prize a solution must be published in a qualifying outlet, survive at least two years of scrutiny and earn general acceptance.[3]
- Alpöge writes that he messaged Dario Amodei late on Sunday and that there was a constraint to respect.[1][3]
- Anthropic began trading on Nasdaq on November 19, the day before the claim.[3]
Implications
- ConfirmedThe claim is public and the proof is posted openly as Lean code.
- Our readingIf it survives checks, it would boost investor confidence in AI companies, Anthropic first of all. This is our inference.
Still unknown
- Whether independent mathematicians accept the proof.
- Whether the Lean statement matches the original hypothesis.
- What timing constraint Alpöge was referring to.
Who it affects
United States, Global. Anyone following AI, tech stock investors
Sources
- Levent Alpöge: Note on a proof of the Riemann Hypothesis November 20, 2026 · official
- Scientific American: Anthropic mathematician claims to prove Riemann hypothesis, most famous problem in all of math November 20, 2026 · media
- Decrypt: Claude Claims Historic Math Breakthrough One Day After Anthropic IPO November 20, 2026 · media
Cover: Freekonomist graphic, a symbolic image, not a photo of the event. This is a news explainer, not financial advice.
More stories
FCC clears 71 companies to sell drones and routers despite the ban
The US has banned new foreign-made drones, routers, robots and inverters. Now 71 companies have an exemption in exchange for US factories.
Gram gold price in Turkey falls to 14-month low as the ounce slides
Gram gold price in Turkey hit a 14-month low on November 20, Haber7 says: 5,845.51 lira (buy). The ounce fell to about $3,500 as a stronger dollar and Fed hike talk weighed on it.
Markets today, November 20, 2026: Nasdaq drops 5% and bitcoin 6% in AI selloff
US stocks, bitcoin and gold fell together on Friday as worries about AI valuations and Fed rate hikes hit risk assets all at once.
Morning brief by email
The night's key moves and what matters today, in 2 minutes. One-click unsubscribe.