The BepiColombo mission is planned to very accurately measure the gravity field, the topography, and the tidal Love numbers of Mercury. In this paper, we review our present knowledge of the interior structure and show how the data from BepiColombo can be used to improve on our knowledge. We show that our present estimates of the core mass and volume depend mostly on our confidence in cosmochemically constrained values of the average silicate shell and core densities. The moment of inertia (MOI) C about the rotation axis will be determined very accurately from the degree 2 components of the gravity field and from measurements of the obliquity and the libration frequency of the rotation axis. The ratio Cm/ C between the MOI of the solid planet to the MOI of the planet, both about the rotation axis, will additionally be obtained. If the core is liquid or if there is a liquid outer core, Cm/ C will be around 0.5. In this case, Cm can be identified with the MOI of the silicate shell. If the core is solid, Cm/ C will be about 1. The MOI C can be used to test and refine present models but will most likely not per se help to increase the confidence in the two-layer model beyond the present level, at least if there is a substantial inner core. C and Cm/ C can be used to calculate the inner core radius and the outer core density, assuming the silicate shell density and the inner core density are given by cosmochemistry. The accuracy of the outer core density estimate depends largely on the confidence in the cosmochemical data. The inner core radius can be determined to the accuracy of the densities if the inner core radius is greater than 0.5 core radii. These values can be checked against the Love number of the planet. The higher order components of the gravity field can be used to estimate core-mantle boundary undulations and crust thickness variations. The former will dominate the gravity field at long wavelength, while the latter will dominate at short wavelengths.