We create a 3-D density, temperature, and composition model of the Antarctic lithosphere using an integrative approach combining gravity, topography, and tomography data with mineral physics constraints and seismic data on crustal structures. The latter is used to create a new Moho and crustal density model. Temperature and thermal density variations are estimated based on S wave velocities from two independent tomography models (SL2013sv and AN1-S). Results of the Antarctic continent show the well-known distinction between East and West Antarctica in temperature and density to a depth of about 200 km. Incorporating compositional variations in the temperature calculations increases temperatures in depleted regions by up to 150°C, giving improved insights into thermal structures. The thickness of the lithospheric root also varies strongly between these regions, with values below 100 km in the west and above 200 km in the east. Regions with negative compositional density variations (<−0.040 g/cm 3 at 100 km), high depletion (Mg # > 91.5), and low temperatures (<800°C; central Dronning Maud Land, along the east flank of the Transantarctic Mountains) are interpreted as Precambrian cratonic fragments. Nearly undepleted lithosphere is found in the Lambert Graben and the Aurora Subglacial Basin and is attributed to Mesozoic rifting activity that has caused lithospheric rejuvenation.Plain Language Summary Antarctica remains one of the least studied areas on Earth, because large ice masses and climate conditions strongly hinder measurements. It plays an important role in global phenomena such as sea level change. In order to understand and predict these processes, we need knowledge about the heat coming from the Earth's interior. It has been recently recognized that the thermal state of the lithosphere, the rigid outer shell of the Earth, plays an important role in controlling dynamics of the ice shield and, thus, global sea level changes. Therefore, we create a model of the lithosphere describing the variation in temperature, density, and composition. Before, such models were created by using some single data set (usually seismic tomography). We employ a new integrative approach, combining several data sets to create a comprehensive 3-D model of the Antarctic lithosphere. Combination of various data sets gives more robust models than when using a single approach. Our results of the Antarctic continent show strong differences between East and West Antarctica in temperature and density to a depth of about 200 km. Regions with negative compositional density variations and low temperatures are identified within East Antarctica and are interpreted as being substantially older than the surrounding continent.