An unreleased Anthropic AI model has made significant progress on the Riemann hypothesis, one of mathematics’ most enduring unsolved problems, the company announced on Monday, August 11, 2026. The hypothesis, which concerns the distribution of prime numbers, has remained unproven for more than 150 years and carries a $1 million prize for a working general proof.
The model did not solve the hypothesis outright, but it substantially increased the lower bound of solutions for which the hypothesis holds true. The result was confirmed by two of Anthropic’s in-house mathematicians and formalized using the open source proof assistant Lean.
What stands out is how the work unfolded. An Anthropic staff member without significant mathematical training prompted the model to “take a real stab” at the problem, then stepped back. Over the following day and a half, the model tested 650 different ideas, coordinated across 60 subagents, and spent 31 million output tokens. Of those subagents, two developed the key mathematical ideas, 13 contributed supporting concepts, 30 attempted but failed to generate new ideas, 13 validated the arguments, and two helped draft the initial paper.
The result is part of a broader wave of AI-driven mathematical discoveries. This year, AI models have solved several Erdős problems, OpenAI’s internal “Astra” model produced 10 major proved results, and a separate Anthropic effort disproved the longstanding Jacobian conjecture.
The trend is generating debate within the mathematics community. In June, a group of prominent mathematicians signed a public declaration raising concerns that AI-generated proofs could undermine the field’s tradition of attributing discoveries to specific authors who take responsibility for their correctness. Fields Medal winner Timothy Gowers offered a different view in a blog post, suggesting AI’s influence could reshape mathematics in more complex and potentially positive ways, comparing the shift to how stars are no longer named after astronomers.
Source: TechCrunch