The accuracy of the Oliver–Pharr approach for nanoindentation experiments critically depends on the interfacial friction condition. Although Coulomb friction is often used in finite element simulations for the correction, the friction stress model may give a more appropriate physical scenario. The measurement of the tangential contact stiffness by a recently developed multidimensional nanocontact system provides a direct verification of these two friction models. Both friction models will predict the tangential stiffness reduction as the consequence of interface microslip, but quantitative comparison to the experiments supports the friction stress model.