Anthropic announced that an unreleased AI model made significant progress on the Riemann hypothesis, a math problem unsolved for over 150 years with a $1 million prize still unclaimed, by substantially raising the lower bound of solutions for which the hypothesis holds true.
Notably, an Anthropic staff member without significant mathematical training prompted the model to attempt the problem, then let it work autonomously for roughly a day and a half. The model tested 650 different approaches, coordinating 60 subagents and using 31 million output tokens, with two subagents responsible for the key mathematical breakthroughs. Two in-house Anthropic mathematicians confirmed the findings, which were formalized using the open-source Lean proof assistant.
The result adds to a growing string of AI-driven mathematical advances in 2026, including AI-solved Erdos problems, OpenAI’s release of ten major results from its internal Astra model, and a separate Anthropic effort disproving the Jacobian conjecture.
These developments have sparked debate within the mathematical community. A June declaration signed by prominent mathematicians warned AI could threaten the field’s tradition of attributing proofs to identifiable authors accountable for their correctness. Fields Medal winner Timothy Gowers offered a different perspective, suggesting mathematics might evolve productively even if theorems become less tied to individual mathematicians, comparing it to how most stars remain unnamed.
