Objective: To measure cuspal flexure of intact and restored maxillary premolars with different restorative materials and occlusal contacts. Method and Materials: Microcomputed tomography data were used to generate finite-element models with a mesioocclusodistal (MOD) cavity preparation and the corresponding restoration. The tooth parts were imported in finite-element software to create 3D volumetric models. Occlusal loading either in enamel, at restoration margin, or in restorative material was simulated by linear contact analysis. Cuspal widening was measured for the different contact locations and restorative conditions (unaltered tooth, MOD porcelain, and composite-inlay restorations). Results: For a given material property, contacts in restoration generated the least amount of cuspal deformation, followed by the enamel contacts and contacts at restoration margin. Models of the intact tooth and ceramic inlay behaved similarly (cuspal widening at 100 N ranging from 1.8 to 3.0 µm and 1.6 to 2.6 µm, respectively). Cuspal flexure was increased in the MOD composite resin restorations (3.4 to 5.7 µm at 100 N of axial force). The least amount of deformation was generated in the ceramic inlay with contacts in only restoration (1.6 µm at 100 N) and the largest in the composite inlay with contacts at restoration margins (5.7 µm at 100 N). For the composite restoration, antagonist contacts at restoration margins were less favorable compared to located on either enamel or the restoration itself. Conclusion: A relatively small cuspal deformation was observed in all models. There is an increased cusp-stabilizing effect of ceramic inlays compared to composite ones. Schlagwörter: cuspal flexure, composite resins, finite element analysis, occlusion, porcelain inlays, restorative dentistry