To resolve controversies over the intrinsic shape of the disaccharide α,α-trehalose and the magnitude of solvent effects, energy surfaces have been computed at different levels of molecular orbital and density functional theory. All quantum mechanical (QM) levels agree that the gauche linkage conformations observed by experimental crystallographic and solution studies are 5-7 kcal/mol lower in energy than that of the trans shape. This is quite different than the findings of the two most recent classical force field studied. In those projects, the trans shape was preferred by 3.3 kcal/mol, and it was inferred that a strong solvent effect was responsible for the gauche experimental conformations. In the QM work, a strong solvent effect is not needed to explain the preference for the gauche form because the trans shapes already have higher energy. A QM continuum model of aqueous solvation has only small effects on the torsional energy surface. The best rationalization of 24 values of the linkage torsion angles from small-molecule crystal structures is provided by HF/6-311++G**//B3LYP/6-31G* theory, a level that under-predicts the strength of hydrogen bonds.