OpenAI reports irrational cubic fourfolds with geometric K3 categories

For every sufficiently large admissible Hassett discriminant, a very general smooth cubic fourfold in that divisor is irrational although its Kuznetsov component is equivalent to the ordinary derived category of a projective K3 surface.

Sources for this record

What was already known

Kuznetsov conjectured that a cubic fourfold is rational exactly when its Kuznetsov component is a geometric K3 category. Known rational examples include special cubic families, and the manuscript notes entire rational divisors C_26, C_38, and C_42 (manuscript introduction, pp. 1–2).

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 Major

Disproves the sufficiency direction of Kuznetsov’s rationality conjecture across all sufficiently large admissible discriminants.

Limitations

The claim is for very general members of each sufficiently large admissible divisor; the discriminant threshold is ineffective and special rational subloci may remain.

Technical verification Reported

Reported is the default when evidence for a higher level is insufficient. The OpenAI-authored manuscript reports and argues the result, but the reviewed sources do not document a completed check of the full assessed claim.

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, and prior-work review for non-notifying local inclusion.

Read this published revision