OpenAI Published 372 Mathematical Proofs
The research results use formal verification to advance work on the Riemann Hypothesis.
Updated on Oct. 7, 2026 in Mathematics
Live Poll
Do you trust artificial intelligence to accurately perform complex scientific and historical research?
OpenAI has released 372 cases of mathematical research, including progress toward solving the Riemann Hypothesis. These proofs utilize the Lean language for logical error verification.
Why it matters
Formal verification of mathematical results using computational reasoning could accelerate proof generation, though the significance of these specific findings remains under peer review.
Each proof required three hours of computational reasoning equivalent to ChatGPT Pro. The results were verified using Lean to ensure logical consistency.
The players
OpenAI
An AI research and deployment company focused on large-scale models and automated reasoning systems.
Carter Church
An engineer who utilized AI to reconstruct historically significant documents.
The details
The system processes formal logic using Lean, an open-source tool for checking mathematical proofs against predefined axioms. In a separate application, the AI model GPT-6 Astra interpreted scanned imagery of a 217-year-old cipher to reconstruct a 1809 military document. By creating a program to analyze character patterns, the AI mapped 24 rows of symbols against partial decryption tables to verify symbol consistency.
Timeline
1809: The original military letter was written.
September 18, 2026: Carter Church restored the military cipher.
October 6, 2026: OpenAI released the mathematical research collection.
The Tech Race
OpenAI is aligning its output with the Lean mathematical theorem prover, a tool central to modern formal proof verification. This move positions the firm alongside academic initiatives that prioritize verifiable logic over the probabilistic outputs characteristic of earlier language models.
These proofs remain in an evaluation phase, meaning they are not yet integrated into standard academic or engineering workflows. Researchers should monitor the coming months for third-party audits of the proofs to confirm their validity before incorporating them into new work.
The takeaway
The intersection of automated reasoning and formal verification could transform how complex conjectures are solved. Watch for the forthcoming academic review of these 372 cases to determine if they hold up against existing literature on the Riemann Hypothesis.
Further reading
For broader developments in automated formal logic, visit our Mathematics section.
Source note: This article includes information reported by 조선일보.
Live Poll
Do you trust artificial intelligence to accurately perform complex scientific and historical research?
