Anthropic Model Advances Riemann Hypothesis Bound

Anthropic said an unreleased AI model increased the lower bound of solutions for which the Riemann hypothesis holds true, reports TechCrunch. The result does not prove the hypothesis, which has remained unsolved for more than 150 years.

The task began when an Anthropic staff member without significant mathematics training prompted the model to try to prove the hypothesis. The model worked for about a day and a half, tested 650 ideas, coordinated 60 subagents, and generated 31 million output tokens.

Two subagents produced the main mathematical ideas. Thirteen contributed ideas to those agents, 30 tried without producing new ideas, 13 checked the arguments, and two helped write the initial paper.

Mathematicians at Anthropic checked the finding. The result was also formalized using Lean, an open source proof assistant.