The dimensional accuracy of digitally processed inlays is often questioned because of inherent manufacturing inconsistencies associated with computer-aided design and manufacturing (CAD/CAM). This study aimed to compare the absolute marginal discrepancy (AMD), marginal gap (MG), internal gap (IG), and overall discrepancy (OD) of three-dimensional (3D) printed, milled and conventional inlays. Forty resin dies were 3D-printed from a class II mesiococclusodistal preparation on a typodont and randomly distributed into four groups of 10 each. Optical impressions were taken for three groups to fabricate CAD/CAM inlays: Group PVC, 3D printed VarseoSmile Crownplus
Group PVT, 3D printed VarseoSmile TriniQ
and Group MVE, milled using Vita Enamic. For Group CGP (control), CGP was conventionally fabricated using Gradia Plus. These inlays were stabilized on dies and subjected to cone‒beam computed tomography to measure discrepancies in mesiodistal and buccolingual sectional images. All the discrepancies differed significantly among the groups (one-way ANOVA, P >
0.05). The mean OD was significantly greater in the MVE than in the PVT and CGP
the mean AMD and MG were significantly greater in the PVC than in the PVT (Tukey test, P >
0.05). Compared with the other groups, the 3D-printed inlays, especially the VarseoSmile TriniQ, presented a closer marginal and internal fit.