The Congo Basin is characterised by a near-circular shape, a pronounced negative free-air gravity anomaly, and a subsidence history that is slow and long-lived. The basin is often considered as an intracratonic basin, implying an unknown formation mechanism. However, the Congo Basin probably initiated by Precambrian rifting and the larger part of its older subsidence history could be explained by post-rift thermal relaxation. The uppermost layer of Mesozoic to Cenozoic sedimentary rocks in the basin appears discontinuous in its evolution and several studies have proposed that these rocks were deposited in response to a process in the mantle. We have examined gravity data and seismic tomographic models to evaluate the role of the sub-crustal mantle in the more recent evolution phase of the Congo Basin. Using seismic tomographic models of the upper mantle and lithospheric thickness models, we show that the Congo Basin is underlain by a thick lithosphere and that the basin boundary likely coincides with the boundary of the Congo Craton. We have reduced the EGM2008 free-air gravity field by correcting for topography and sediments. We find that the observed negative gravity anomaly is mainly due to the sedimentary units in the basin. The reduced gravity field has slightly negative to positive anomalies over the basin, depending on the densities assigned to the sedimentary rock package. We have analysed thirteen whole-mantle and five upper-mantle tomographic models and show that they do not provide supporting evidence that the sub-lithospheric mantle played a primary role in the more recent subsidence of the Congo Basin. We speculate that deposition of the Mesozoic–Cenozoic rocks could have raised the surface elevation of the Congo Basin to the present average level of ∼400 m above sea-level and that the last subsidence phase could be a consequence of the sediment load rather than the cause.