VibeMathedMath problems solved with AI

Cycle-residue stability at minimum degree five

Classify the finite simple graphs of minimum degree at least five whose cycle lengths fail to represent every residue class modulo five. Is residue two the only possible missing residue, and can all such graphs be characterized through an explicit family of exceptional end-blocks together with a condition on the remaining blocks?

Result
Proved(see note)
Status
Candidate (review pending)
AI contribution
AI co-developed
Method
Argument
Field
Graph Theory
Posed by
Luo, Ma and Zhao, whose stability theorem covers every k >= 6 and leaves k = 5
Year posed
2026
Years open
0y
Solved
2026-09-04
Model
GPT-5.6 Sol, GPT-6 Astra
Vendor
OpenAI
Collaborators
Elias Botsford
Verification
Unreviewed
Publication
Preprint
Significance
20 / 100
Disclosed cost
Wikipedia
No dedicated article

What was actually shown

The preprint claims an exact classification. Write C5(G)C_5(G) for the residues modulo five represented by cycle lengths in GG. Let
E5=K6,K5,5H5,n;t:2t5<n\mathcal E_5={K_6,K_{5,5}}\cup{H_{5,n;t}:2\le t\le5<n},
where H5,n;tH_{5,n;t} is obtained from K5,nK_{5,n} by deleting 5t5-t edges incident with one vertex in the part of size nn.

For every finite simple graph GG with minimum degree at least five, exactly one alternative holds: C5(G)=Z5C_5(G)=\mathbb Z_5; or every end-block belongs to E5\mathcal E_5 and every non-end-block contains no cycle of length congruent to two modulo five. Every member of E5\mathcal E_5 has cycle-residue spectrum 0,1,3,4{0,1,3,4}.

The proof combines structural arguments with finite computational checks. It uses the separately established Dean–5 theorem and its weak-graph strengthening as inputs. The contribution is the stronger stability classification; independent expert review remains pending.

What the AI did

Large language models contributed substantially to developing and auditing the proof, repairing intermediate arguments, and implementing computational checks. The work included checking rooted-path and cycle constructions, identifying gaps in structural reductions, developing replacement lemmas and corrected attachment arguments, and revising finite residue calculations and literal-cycle geometry checks.

Verification

Unreviewed. The 82-page preprint (Zenodo 10.5281/zenodo.22311785, version 1.0.0 of 4 September 2026) was text-extracted and its introduction, main theorem and disclosure read here; the computational supplement (10.5281/zenodo.22311412) reports 23 passing operations and was not replayed on this site, unlike the Dean-5 supplement, which was. Two dependencies a reader should hold in mind: the proof imports the author's own modulus-five Dean theorem, version 1.0.1, which is itself Candidate here pending independent review; and the earlier Zenodo version of this preprint was withdrawn by the author for errors before this one, which is stated in the submission and matches the deleted record. Tier changed from AI-discovered to AI co-developed: the paper's own disclosure says the results were obtained "with substantial assistance from large language models" and that the author reviewed the arguments and takes responsibility, which is the co-developed pattern on this site, not discovery.

Sources

Submitted by eli on

Changelog2 changes

Discussion