Oda’s smoothness conjecture
For every full-dimensional smooth lattice polytope and every integer , every lattice point can be written as for some lattice points . Here, smooth means that at each vertex of , the primitive edge-direction vectors form a -basis of .
References
- Integer decomposition property of three-dimensional smooth polytopes: a shorter proof — arXiv — Juliana Curtis, Gaku Liu
Source: arXiv
Progress summary
A new paper claims to prove the conjecture in three dimensions, but the general conjecture remains open and the proof has not been independently checked.
Oda posed the conjecture at Oberwolfach in 1997: every smooth lattice polytope should have the integer decomposition property. The new claim concerns only the three-dimensional specialization with , not the stronger formulation for arbitrary pairs.
Known results
- The two-dimensional case is known (Fakhruddin; independent proof by Ogata).
- Centrally symmetric three-dimensional polytopes satisfy the property (2018 paper, published 2019).
- Smooth polytopes with at most facets satisfy it (Robins, submitted 2021; revised 2024).
- Further special cases include smooth combinatorial cubes in every dimension and three-dimensional pseudo-prisms (Curtis, 2025).
October 12, 2026 claimed three-dimensional proof
Juliana Curtis and Gaku Liu presented a shorter proof of the claimed three-dimensional result; Mateusz Michałek’s related preprint claims normality for every smooth lattice three-polytope by proving -normality. It explicitly covers only . Michałek says the decisive part was achieved by OpenAI Sol Ultra in Codex, with other named AI systems used for checking and writing. Neither proof has independent expert validation.
Current status (as of October 2026): The two-dimensional case and several special cases are known; a three-dimensional proof is claimed but unverified, and the full conjecture remains open.
Solutions 0
No solutions have been posted yet.