Transcatheter aortic valve implantation shortcomings such as relevant paravalvular leakage (PVL) have been shown to negatively impact survival. The Sapien 3 prosthesis has added features to address these problems. We compared the functional performance of the Sapien 3 (S3) and its predecessor Sapien XT (XT), with a particular focus on PVL. We analyzed 601 patients presenting with severe native aortic valve stenosis treated with either Sapien XT (n = 405, 2011 to 2014, mean STS-PROM [Society of Thoracic Surgeons predicted risk of mortality] 11.0%) or Sapien 3 prosthesis (n = 196, 2014 to 2016, mean STS-PROM 9.2%). Beside a propensity score–based comparison, we modeled the likelihood of PVL as a function of oversizing and prosthesis. Primary end points were 1-year survival, degree of oversizing, and occurrence of PVL. One-year survival (stratified log-rank p = 1.00) and 30-day mortality (S3: 10 of 126 vs XT: 10 of 126, p = 1.00) did not differ. In the matched cohort, oversizing was less common in the S3 group (absolute median difference of 7% in oversizing, interquartile range 1.1% to 12.7%, p = 0.025). PVL > = 1° was similar in both groups (S3: 13 of 126 vs XT: 20 of 126, p = 0.296). Mean gradients were lower in the XT group (median difference 1.0 mm Hg, interquartile range 0.3 to 1.8 mm Hg, p = 0.005). Rate of postdilatation, implantation of a second valve (valve-in-valve), annular rupture, and new pacemaker implantation were similar (all p = 1.00). In conclusion, compared with the Sapien XT, the redesigned Sapien 3 prosthesis offers effective sealing against PVL without grossly compromising hemodynamic performance or increasing the necessity for new pacemaker implantation.