Breaking

OpenAI AI System Claims to Disprove Exponential Time Hypothesis, Expert Review Underway

October 27, 2026 at 10:02 EDT

Foundation Models
Research & Papers

OpenAI announced on October 26, 2026 that one of its internal AI models has produced a mathematical proof refuting the Exponential Time Hypothesis (ETH), a foundational 1999 conjecture in computational complexity theory, though independent verification by leading researchers is still underway.

OCTOBER 26, 2026 · AI RESEARCH

AI Disproves 27-Year Computing Conjecture

OpenAI's system produces claimed proof that ETH is false, undergoes expert verification by leading complexity theorists.

Timeline

ETH Proposed

1999

Natural Proof Found

Oct 16

Lean Formalization

Oct 17

Public Announcement

Oct 26

The Verification Process

Leadership

Russell Impagliazzo (UC San Diego)—ETH co-creator
Ramamohan Paturi (UC San Diego)—ETH co-creator
Ryan Williams (MIT)—Complexity theorist

Key Conditions

No nondisclosure agreements signed
No payment to reviewers (unpaid work)
All raw data published publicly
OpenAI $5M contribution to research

What the Result Does & Doesn't Mean

Refutes ETH: Algorithm asymptotically beats 2^(o(n))

Does Not Solve: P vs. NP remains open

Machine-Certified: Lean formal verification complete

Not Practical: No real-world speedup for instances

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.