Status on AI Efforts to Prove Reimann Hypothesis (Google Gemini)

Search instead for is ai any closer to solving the rheimann sum problem

AI Overview

Artificial intelligence has not solved the Riemann Hypothesis, but an unreleased research model of Anthropic's Claude made a notable mathematical advance on a related sub-problem in August 2026. 

(Note: While the prompt mentions the "Riemann sum"—a basic calculus method for approximating areas under curves—this recent breakthrough and the $1 million prize apply to the advanced Riemann Hypothesis about prime numbers.)

Recent AI Progress

The Breakthrough: In August 2026, Anthropic announced that an unreleased version of Claude improved a specific numerical bound related to the distribution of prime numbers (the zero-free region estimate). 
The Numbers: Claude pushed the lower bound of this estimate up from 41.6% to 67.2%. 
How It Worked: An Anthropic employee named Jared Sumner prompted the AI and provided steady, motivational encouragement while the model ran a massive, multi-agent numerical search through thousands of calculations and failed ideas. 
Mathematicians' Reactions: Experts like Oxford mathematician James Maynard noted that the result provides a genuinely interesting mathematical contribution and a useful "workaround," but it is not a direct path to solving the core hypothesis. 

Current Status of the Problem

Still Unsolved: The core hypothesis—stating that all non-trivial zeros of the Riemann zeta function lie on the critical line—remains completely unproven.
The Prize: The $1 million Millennium Prize offered for the official proof remains unclaimed.
The Limit of AI: Experts emphasize that running massive parallel computer searches for better numbers is very different from creating the deep, intuitive theoretical insights needed for a true mathematical proof. 

If you'd like to dive deeper into this topic, I can help you:

• Explain the Riemann Hypothesis in simple terms
• Explore how AI assists in other math problems (like Navier-Stokes or geometry)
• Break down how prime numbers connect to this equation

Let me know what you would like to explore next.

OpenAI Agents Have Solved (?) Navier-Stokes Problem (Google AI Response)

 

OpenAI announced that its AI agents solved a portion of the Navier-Stokes existence and smoothness problem, one of the seven prestigious Millennium Prize Problems worth $1 million. [1, 2]

The Breakthrough

What was solved: An analytical proof showing that a singularity (a breakdown or infinite growth in fluid velocity) can form in a three-dimensional fluid motion under a smooth external force, resolving statements "C" and "D" of the official Clay Institute problem formulation. [1, 2, 3]
How it was done: OpenAI deployed a large group of roughly 10,000 autonomous AI agents working together using a new unreleased internal model. [1, 2]
Time and cost: The agents reached the solution in 88 hours (under 4 days), exchanging nearly 5 million messages and consuming roughly 130 billion output tokens with an estimated compute cost of over $10 million. [1, 2, 3]
The prize: OpenAI stated they do not intend to claim the $1 million reward, viewing the achievement as a demonstration of technical capability rather than an official prize submission. [1, 2]

Controversy and Context

Dispute over timing: The announcement triggered a debate after it was revealed that human researchers had been circulating preliminary or unpublished notes on similar approaches right before the AI run, leading some mathematicians to question the originality or practical significance of using a contrived external force. [1, 2]
Mathematical reaction: While some experts note the solution fits the literal wording of the Clay Institute's option "C", others argue it relies on unnatural real-world constraints and doesn't answer the unforced turbulence problem mathematicians ultimately care about. [1]

A news report discussing the details and reactions to OpenAI's AI system cracking the Navier-Stokes problem: