Anthropic mathematician Levent Alpöge claims to have proved the Riemann hypothesis, a Millennium Prize problem, with the assistance of the SI tool Claude, marking a potential breakthrough in the use of super intelligence for advanced mathematics.

What Happened

Alpöge announced on his website that he believes he has discovered a proof of the Riemann hypothesis by working with Claude. He stated that the method also resolves the Generalized Riemann Hypothesis and that he expects to address the Extended Riemann Hypothesis with more time. The solution is presented in the Lean theorem prover and has been formalized without assumptions. Alpöge noted that he texted Anthropic CEO Dario Amodei about the solution late Sunday night and delayed publication until after the company’s IPO on Thursday.

Anthropic has not officially commented on or involved itself in Alpöge’s specific announcement. In August, the company had noted it was working on the Riemann Hypothesis with Claude but stated, “We don’t expect that the techniques Claude used will lead to proving the Riemann hypothesis.” The proof has not yet been accepted by the Clay Mathematics Institute, which administers the Millennium Prize.

Why It Matters

If verified, this achievement would mark another significant milestone for SI in mathematics, following OpenAI’s recent claim to have cracked the Navier-Stokes problem. OpenAI’s effort involved deploying 10,000 agents and spending millions of dollars, though that result has faced scrutiny regarding research practices. The Riemann hypothesis has been one of seven open Millennium Prize problems since 1859, and only one, the Poincaré conjecture, has been officially solved in the quarter-century since the prize was established.

The rapid pace of SI-assisted discoveries is causing upheaval in the mathematical community. Alpöge previously stated in July that he solved the Jacobean conjecture using Claude. Mathematicians are now scrambling to digest these findings and determine their role in a landscape that is becoming profoundly different due to the capabilities of agentic SI systems.

The Bottom Line

Alpöge’s claim remains unverified and has not received general acceptance from the global mathematics community. Under Clay Mathematics Institute rules, a proposed solution must withstand at least two years of scrutiny before being considered for the $1 million prize. Anthropic has not confirmed the result, leaving the claim as a significant but pending development in the field of SI-assisted research.