🔍 Read the full analysis: OpenAI’s AI Didn’t Prove Everything The Headline Suggests on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
OpenAI says an unreleased AI model produced a proof that a deterministic algorithm can exactly count solutions to fixed-width SAT formulas in subexponential time, contradicting ETH for 3-SAT. Lean checked a formalized version with a weaker runtime bound than the preprint’s sharpest result, and independent expert review is still pending; the claim does not settle P versus NP or promise immediate practical speedups.
OpenAI says an unreleased AI model has produced a proof of a subexponential-time algorithm for exactly counting solutions to fixed-width SAT formulas, a result that would refute the Exponential Time Hypothesis (ETH) if correct. The company released a preprint and a Lean formalization, but independent expert review is still pending, and the formalized statement checks a weaker runtime bound than the paper’s sharpest claim.
The preprint’s central claim is that a deterministic algorithm can exactly count satisfying assignments for any k-SAT formula with fixed k in subexponential time and polynomial memory. For 3-SAT, that would contradict ETH, which says no algorithm can solve the problem in subexponential time. The paper says the result also refutes the stronger Strong Exponential Time Hypothesis and three related hypotheses.
OpenAI says its system found the disproof on October 16, with the Lean formalization following on October 17. The manuscript is dated October 17 and was revised October 25. The source report says the public Lean release contains 3,977 files and about 710,000 lines, based on its author’s count. Lean’s machine check supports the formalized proof statement; it does not amount to independent expert endorsement of every claim in the paper.
A key qualification is the gap between the paper and the formalization: according to the source report, Lean checks a slightly weaker runtime bound. The preprint’s sharpest bound is not covered by the formalization. A group led by Russell Impagliazzo, Ramamohan Paturi and Ryan Williams is expected to examine the computer-science result. Impagliazzo and Paturi were among the original authors of ETH.
THREE RESULTS, THREE RELEASE STRATEGIES
What OpenAI released, what Lean checked, who reviews it
Dates and the physics status are OpenAI's own account. Lean-checked means a machine verified the proof, not that experts have accepted it.
Local smoothing, so Kakeya too
- AI output
- 18 Oct 2026
- Lean
- 19 Oct 2026 · strict ranges
- Public
- AI write-up + Lean proof
- Review
- No group: researchers organize themselves
The Exponential Time Hypothesis is false
- AI output
- 16 Oct 2026
- Lean
- 17 Oct 2026 · weaker bound
- Public
- 67-page preprint + Lean proof
- Review
- Impagliazzo, Paturi (UC San Diego), Williams (MIT)
QCD exists; pure Yang–Mills as sub-result
- AI output
- 9 Sep (Yang–Mills), 18 Sep (QCD)
- Lean
- Yang–Mills existence, 16 Oct (OpenAI)
- Public
- No: the group holds the AI materials back
- Review
- Gross, Douglas, Jaffe, Klebanov, Nekrasov
Before this release
The repository the results sit in
- 6 Oct: more than 700 AI-generated manuscripts published
- 7 Oct: 3 withdrawals after a sign error; 14 other manuscripts revised
- About 42% of top-line results formalized in Lean (repository’s own figure)
- OpenAI calls the manuscripts “raw research data”, not polished papers
Sources: OpenAI, 26 Oct 2026 · Lean note 378 · Lean note 079 · repository history. Status chips: our assessment.
Powered by Thorsten Meyer AI
A Challenge to SAT Hardness
ETH has been used to support conditional lower bounds across theoretical computer science: researchers use it to argue that certain algorithms are unlikely to run substantially faster than known methods. If the proof survives scrutiny, those conclusions may need to be revisited. The result would challenge a major complexity-theory hypothesis, not show that existing software or everyday computing will suddenly become faster.
The claimed algorithm runs in roughly 2^(n / √log n) time, according to the source material. That is subexponential, but it is still far slower than polynomial time. OpenAI said practical improvements “will require significant further research.” The near-term significance is chiefly theoretical: whether a long-standing hardness assumption can be disproved, and whether AI-produced mathematics can be checked and evaluated through established formal methods.
The result is also not a solution to P versus NP. ETH is a stronger assumption about the difficulty of solving SAT. Refuting ETH would not establish that every efficiently checkable problem is efficiently solvable, and P versus NP remains open.
Products Worth Considering
LG UltraGear 27-Inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDMI 2.1
DJI Osmo Mobile 8 Gimbal with Mic Mini Transmitter, Infinity Black
Logitech MX Master 4 for Mac, Ergonomic Bluetooth Mouse with Haptic Feedback, Ultra-Fast Scrolling, USB-C Charging, for MacOS, iPadOS, Windo
As an affiliate, we earn on qualifying purchases.
From SAT Search to ETH
SAT asks whether truth values can be assigned to variables so that every logical clause is satisfied. In k-SAT, each clause contains at most k variables or their negations. A direct search over n variables considers 2^n assignments, though decades of work have improved on brute force by reducing the exponential base.
ETH, proposed by Russell Impagliazzo, Ramamohan Paturi and Francis Zane in 2001, makes a stronger claim: 3-SAT cannot be solved in subexponential time. It is distinct from P versus NP, which asks whether every problem whose proposed answers can be checked efficiently can also be solved efficiently. The Clay Mathematics Institute still lists P versus NP as unsolved.
The ETH announcement was one of three research claims OpenAI described on October 26. The others concern a proof in harmonic analysis and claims about quantum chromodynamics and a grand unified completion of the Standard Model. OpenAI says outside researchers are shepherding the releases; the physics materials are being held back at the group’s request. The source report also notes that OpenAI’s earlier October 6 release of more than 700 AI-generated manuscripts was followed by withdrawals and revisions, underscoring why formal checks and outside review matter.
Expert Review and Formal Scope
The central unresolved issue is whether independent experts will validate the full argument and its implications. The review group’s examination is pending, and the source material does not report a final assessment. A machine-checked formalization can catch errors in the statement and proof encoded in Lean, but it does not by itself establish that the formalization covers every detail or the strongest runtime bound in the preprint.
It is also not yet clear whether the result will lead to algorithms that improve practical SAT solving or other applications. The claimed asymptotic bound alone does not answer that question, and OpenAI has said further research is needed. The source report gives no confirmed timetable for the review’s conclusion.
Review Before Wider Claims
The next milestone is the independent examination by the group led by Impagliazzo, Paturi and Williams. Its assessment could clarify whether the preprint’s proof holds, how the gap between the paper’s sharpest bound and the Lean statement should be understood, and what conclusions researchers can safely draw.
Until that review is reported, the result should be described as OpenAI’s claim, supported by a public preprint and a machine-checked formalization of a weaker bound, rather than as a settled refutation of ETH. Any practical consequences would require additional research and are not established by the announcement.
Key Questions
What does OpenAI say its AI proved?
OpenAI says an AI model produced a proof of a deterministic, subexponential-time algorithm for exactly counting solutions to fixed-width k-SAT formulas. Applied to 3-SAT, the claim contradicts ETH if the proof is valid.
Has the result been independently verified?
Not yet, according to the source material. A Lean formalization has been released and machine-checked, while an external group is expected to examine the result. The sharpest runtime bound in the preprint is not included in the formalized statement.
Does this solve P versus NP?
No. The claim concerns ETH, a stronger hypothesis about the difficulty of SAT. P versus NP remains open, and the claimed algorithm is still superpolynomial.
Will SAT programs become faster now?
That has not been established. The source report says the claimed asymptotic algorithm is not a practical speedup by itself, and OpenAI said practical improvements would require significant further research.
Source: ThorstenMeyerAI.com
