OpenAI reports a depth-three circuit lower bound beyond every square-root constant

One fixed polynomial-time Boolean language has exact depth-three OR–AND–OR circuit size greater than 2^(A√n) for every fixed A>0 at all sufficiently large input lengths n.

Sources for this record

What was already known

The manuscript describes classical explicit-function depth-three lower bounds with an exponent proportional to √n and identifies the question of exceeding every such constant for one fixed language.

AI involvement AI-led

OpenAI credits itself as manuscript author and states that its internal frontier model produced the results in this release. Individual human input is not detailed enough to establish task-only autonomy.

OpenAI is named as author on the manuscript, and its release attributes the results to an internal frontier model. The individual human input is not detailed enough to establish autonomy.

Importance Notable

Breaks the square-root-constant barrier for unrestricted-fan-in depth-three Boolean circuits.

Limitations

The result does not give a fixed exponent n^(1/2+ε), nor a lower bound for unrestricted circuit depth.

Technical verification Formal (Lean 4)

OpenAI reports a Lean proof matching the selected claim.

Formal checking

The pinned OpenAI Lean scope note identifies this selected claim.

Checker
OpenAI-reported Lean 4 / Comparator
Scope
One fixed polynomial-time Boolean language has exact depth-three OR–AND–OR circuit size greater than 2^(A√n) for every fixed A>0 at all sufficiently large input lengths n.
Basis
Reported by the linked checker
Evidence source

Checking a formal statement does not establish that it matches the original problem.

Human understanding Few

Few is the default when evidence for a higher level is insufficient. No named person is documented as explaining the central argument for this assessed contribution.

Correction and publication history

2026-10-18 · publication

Primary manuscript, attribution, prior-work, and scoped checking review for non-notifying local inclusion.

Read this published revision