Formation of Terrestrial Planets

International audience; The past decade has seen major progress in our understanding of terrestrial planet formation. Yet key questions remain. In this review we first address the growth of 100 km-scale planetesimals as a consequence of dust coagulation and concentration, with current models favoring the streaming instability. Planetesimals grow into Mars-sized (or larger) planetary embryos by a combination of pebble- and planetesimal accretion. Models for the final assembly of the inner Solar System must match constraints related to the terrestrial planets and asteroids including their orbital and compositional distributions and inferred growth timescales. Two current models -- the Grand-Tack and low-mass (or empty) primordial asteroid belt scenarios -- can each match the empirical constraints but both have key uncertainties that require further study. We present formation models for close-in super-Earths -- the closest current analogs to our own terrestrial planets despite their very different formation histories -- and for terrestrial exoplanets in gas giant systems. We explain why super-Earth systems cannot form in-situ but rather may be the result of inward gas-driven migration followed by the disruption of compact resonant chains. The Solar System is unlikely to have harbored an early system of super-Earths; rather, Jupiter's early formation may have blocked the ice giants' inward migration. Finally, we present a chain of events that may explain why our Solar System looks different than more than 99\% of exoplanet systems.