Claude Helps Prove Riemann Hypothesis: 250k-Page Lean Formalization Awaits Verification
Levent Alpöge claims he and Claude have proven the Riemann Hypothesis, formalizing a 250,000-page proof in Lean that may extend to the Generalized Riemann Hypothesis, but independent verification of 175,187 modules is ongoing and the result is not yet accepted by mathematicians.
Claimed Proof of the Riemann Hypothesis with AI Assistance
On November 20, mathematician Levent Alpöge posted a four-page note on his personal website stating that he and the AI system Claude have proven the Riemann Hypothesis. The proof is fully formalized in the Lean theorem prover without additional assumptions, and Alpöge suggests the method can be extended to the Generalized Riemann Hypothesis, with the Extended Riemann Hypothesis requiring more time.
Scale and Nature of the Formalization
The argument expands to over 250,000 pages — a volume Alpöge describes as a "book" that no human could read. Consequently, he released only the Lean formalization code, split into 14 compressed archives (linked in the note's footnotes), rather than a human-readable paper. Alpöge admits he has not fully internalized the vast mathematical landscape involved; he had planned to spend decades on the problem, but Claude raced ahead at the finish line, in his words "fluttering happily and brightly" near the endpoint.
Key Technical Insight
Alpöge highlights a specific step that would shock number theorists: for the von Mangoldt explicit formula, Claude proposed using a sequence of θ lifts to mimic Stepanov's proof in the function field setting, with chain length set to √X. In function fields, Stepanov's method relies on a divisor with multiplicity; over the integers no such object exists, and constructing an analogue is akin to differentiating integers — a point Alpöge underscores with the remark "otherwise how long could it be?".
Timeline and Disclosure
The note includes a timeline compiled afterward by Claude, with the final entry dated November 16 marking the Generalized Riemann Hypothesis. Alpöge cautions that for parts where Claude understands more than he does, he is merely relaying a summary that may be incomprehensible, and the Lean code remains the authoritative source. On the night before the announcement, Alpöge messaged Dario Amodei (CEO of Anthropic) asking permission to publish, acknowledging a mandatory restriction and poor timing but feeling he could not wait longer.
Independent Verification Underway
A volunteer-initiated GitHub repository ( https://github.com/rohanarun/alpoge-rh-replay) has begun re-running the Lean code. As of the time of writing, the full re-check is still in progress across 175,187 modules, and no verification receipt for the Riemann Hypothesis has been issued. This effort does not constitute acceptance by the mathematical community.
Source References
Alpöge's note: https://alpo.ge/remark.pdf Verification repository:
https://github.com/rohanarun/alpoge-rh-replaySigned-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
