Dialysis tubing (and most semi-permeable membranes) operate based on differences of size and not molecular weight. While the size of a molecule or ion does increase as its molecular weight increases, there is some nonintuitive size-mass discrepancies that can occur.
Many of these discrepancies are well known to polymer chemists. Most of the ways that we attempt to measure the average molecular weight of a polymer sample are actually measures of average molecular size. The shape that a molecule adopts in solution greatly impacts its apparent size. A polymer of MW in the 40 kDa range that adopts a rigid-rod conformation in solution will usually have a larger apparent size than a polymer of MW in the 100 kDa range that folds up into a tight globular conformation (spheres have a smaller surface area to volume ratio than cylinders).
Another factor in the "size" of a molecule is the number of solvent molecules that get dragged along with it. If the molecule is dissolved in a good solvent (for example sucrose in water), it will always have a number os solvent molecules associated with it. The solvodynamic radius of a molecule may be much greater that the molecule's actual size.
If indeed you had dialysis tubing rated to retain only greater than 12,000 Da (and like @F'x, I'm not sure that you did) it is still possible for sucrose to not pass through if the sucrose is trying to drag too many water molecules (from hydrogen bonding) with it.