---
title: "Anthropic AI Model Shows Progress on 150-Year-Old Riemann Hypothesis"
url: https://projectchintan.com/article/anthropic-ai-progress-riemann-hypothesis-math-problem-hlyjk
publisher: Project Chintan
author: Project Chintan Newsroom
section: Technology
published: 2026-08-11T17:59:10.672Z
modified: 2026-08-11T21:30:04.668Z
language: en-IN
---

# Anthropic AI Model Shows Progress on 150-Year-Old Riemann Hypothesis

An unreleased Anthropic AI model has demonstrated significant progress on the Riemann hypothesis, a complex mathematical problem unsolved for over 150 years. The AI explored 650 ideas, coordinating 60 sub-agents over 31 million computational units to advance potential solutions.

## Key takeaways

- An unreleased Anthropic AI model has significantly advanced work on the 150-year-old Riemann hypothesis.
- The AI explored 650 mathematical ideas using 60 sub-agents and millions of computational units.
- This development is part of a trend of AI models making breakthroughs in mathematical problem-solving.
- The progress has raised discussions in the mathematics field about AI's role in discovery and authorship.

An artificial intelligence system developed by Anthropic has achieved notable advancements concerning the Riemann hypothesis, a mathematical problem that has remained unsolved for more than a century. The hypothesis relates to the distribution of prime numbers and carries a $1 million bounty for a verified general proof.

The Anthropic model, which has not yet been released, substantially increased the lower bound for which the hypothesis holds true. This progress was facilitated by an Anthropic employee with limited mathematical expertise, who instructed the AI to attempt a proof. Over approximately 36 hours, the AI system, comprising 60 sub-agents, explored 650 distinct strategies, consuming a total of 31 million computational units.

## Key Facts

- An unreleased Anthropic model has made significant progress on the Riemann hypothesis.
- The Riemann hypothesis is a long-standing unsolved problem in mathematics concerning prime number distribution.
- A $1 million bounty exists for a general proof of the hypothesis.
- The Anthropic model tested 650 different ideas to address the problem.
- The AI utilized 60 sub-agents and a total of 31 million computational units.
- Two sub-agents were identified as developing the key mathematical ideas.
- The findings were confirmed by Anthropic's in-house mathematicians and formalized using the Lean proof assistant.

## Background

This development follows a series of mathematical achievements attributed to Large Language Models (LLMs). Earlier this year, AI models have reportedly solved various Erdos problems. OpenAI's "Astra" model has produced ten significant results, and another Anthropic effort disproved the Jacobian conjecture.

These advances have sparked debate within the mathematical community. In June, a group of prominent mathematicians issued a statement expressing concern that AI could diminish the value of human authorship and accountability in mathematical discovery. Conversely, some experts, like Fields Medal winner Timothy Gowers, suggest that AI's role may lead to a more complex and potentially positive transformation of mathematics, likening it to the historical naming conventions of stars.

---
Canonical: https://projectchintan.com/article/anthropic-ai-progress-riemann-hypothesis-math-problem-hlyjk
Reported from: Multiple Sources