Matiyasevich's single-fold conjecture: every r.e. set has a single-fold Diophantine representation
By the Davis-Putnam-Robinson-Matiyasevich theorem every recursively enumerable set has a Diophantine representation: a polynomial with integer coefficients such that iff for some . The representation is single-fold if such is unique when it exists, finite-fold if there are finitely many. Matiyasevich (1974) proved single-fold exponential Diophantine representations and posed the polynomial case, reducing it to a single-fold representation of exponentiation; the finite-fold conjecture is the weaker form. Does every recursively enumerable set have a single-fold Diophantine representation?
- Result
- Proved(see note)
- Status
- Candidate (review pending)
- AI contribution
- AI-discovered
- Method
- Argument
- Field
- Computability; Hilbert's tenth problem
- Posed by
- Yuri Matiyasevich
- Year posed
- 1974
- Years open
- 52y
- Solved
- 2026-09-24
- Model
- Unreleased internal OpenAI model
- Vendor
- OpenAI
- Collaborators
- —
- Verification
- Lean-checked, statement unaudited
- Publication
- Announced
- Collection
- OpenAI math release (October 2026), version adc7f12
- Significance
- 42 / 100
- Disclosed cost
- —
- Wikipedia
- No dedicated article
What was actually shown
Theorem 1.1: every r.e. , , has a polynomial with equal to 1 for and 0 otherwise. Hence the finite-fold conjecture holds, and Diophantine solvability over stays undecidable under an at-most-one-solution promise (Section 8). The construction bounds a Pell-type exponential with a cutoff from rational points on a rank-one elliptic curve. Witnesses range over ; the paper does not claim the analogue over beyond what follows from it, nor any bound on degree or number of variables.
What the AI did
The release README says every result in it was produced by an unreleased internal OpenAI model following a fixed procedure, on average about three hours of ChatGPT Pro thinking compute per result, and that some outputs build on earlier model results. This result is not among the README's exceptions (the Hodge conjecture for CM abelian varieties and the Re(s) > 11/12 zero-free region). The manuscript is authored 'OpenAI' and names no human author. Single-manuscript family dated September 24, 2026.
Verification
No independent mathematician has checked this yet. Checked here: Theorem 1.1 was read against Matiyasevich's conjecture; it counts all auxiliary variables and gives exactly one complete witness tuple for members and none for nonmembers, over . lean/formalization.yaml lists a main result (comparator SingleFold, declaration OAI.SingleFold.main). The comparator statement was read here: for every and every r.e. (via Mathlib's REPred) there are and an integer polynomial in variables with iff some is a zero, and the zero unique for each . This states the headline. The undecidability-under-promise corollary is not separately formalized (lean/docs/242.md). Not rebuilt here.