ObjectivesTo test strength and reliability of 3D printed compared to milled zirconia. MethodsCylindrical specimens were fabricated from milled (group G1; e.max ZirCAD LT) and from 3D printed (group G2; LithaCon 3Y 230) 3-mol% yttria-stabilized tetragonal zirconia polycrystal (3Y-TZP). While G1 and G2 were sintered in one step, a further series (G3) of 3D printed 3Y-TZP was sintered in two steps including intermediate color infiltration. In each group, two different conditioning strategies were applied (n ≥ 20 samples/subgroup): (1) final polishing with #1200 diamond discs according to ISO 6872, and (2) final polishing with #220 diamond discs resulting in imperfectly polished surfaces. All samples were tested to fracture with a universal testing device (cross-head speed: 1 mm/min). Characteristic strengths and Weibull moduli were calculated. Effects were analyzed by means of either ANOVA (homocedastic data) or Welch ANOVA (heterocedastic data). ResultsFor samples conditioned according to ISO 6872, mean flexural strengths were 1462 ± 105 MPa (G1), 1369 ± 280 MPa (G2), and 1197 ± 317 MPa (G3). For the imperfectly polished subgroups, strength values were 1461 ± 121 MPa (G1), 1349 ± 332 MPa (G2), and 1271 ± 272 MPa (G3). Although all groups showed high mean strength values, the reliability of milled zirconia (Weibull moduli 14 < m <16) outperformed that of the 3D-printed material (3 < m <6). SignificanceEven after color infiltration in a partially sintered state, the tested 3D printed zirconia exceeded the ISO flexural strength criteria for all types of fixed ceramic restorations by far (800 MPa for class 6, ISO 6872), indicating its high potential for clinical use. Further optimization of the internal material structure after sintering might improve the reliability of 3D printed zirconia which is currently inferior to that of milled zirconia.