Unreleased New Model "Shines," Anthropic Announces Claude Makes Major Breakthrough on Riemann Hypothesis

Anthropic announced that its research version of the Claude model has made significant progress in the field of mathematics. In autonomous testing, the model successfully raised the long-term lower bound of the proportion of zeros on the critical line of the Riemann zeta function from 41.6% to 67.2%. This achievement demonstrates the reasoning and computational capabilities of large models on complex mathematical problems.

The key to this breakthrough lies in the model's autonomous execution capability. Running in the Claude Code environment, Claude coordinated 60 sub-agents after experiencing 650 failed attempts. It cumulatively executed 2,400 Shell commands and wrote hundreds of Python scripts, demonstrating the potential of agents in planning and executing long-chain tasks.

This event marks a new step for the application of AI in scientific research. Traditionally, mathematical conjectures have relied mainly on the intuition and proofs of human mathematicians, while the intervention of AI provides new problem-solving paths. Through large-scale computation and automated scripts, the model can explore possibility spaces that are difficult for humans to exhaust.

As a major problem in the mathematical world, progress in the research of the Riemann Hypothesis has profound significance. The performance of AI models on such problems predicts the maturity of the future "AI for Science" model. Large models are expected to become powerful assistants for researchers, accelerating discoveries and breakthroughs in the field of basic science.

This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.