Sorafenib is a second-generation, orally active multikinase inhibitor that is approved for the treatment of patients with advanced renal cell carcinoma and patients with unresectable hepatocellular carcinoma. We studied active transport of sorafenib in MDCK-II cells expressing human P-glycoprotein (P-gp/ABCB1) or ABCG2 (breast cancer resistance protein) or murine Abcg2. Sorafenib was moderately transported by P-gp and more efficiently by ABCG2 and Abcg2. Because sorafenib is taken orally, we orally administered sorafenib to wild-type, Abcb1a/1b(-/-), Abcg2(-/-), and Abcb1a/1b;Abcg2(-/-) mice, completely lacking functional Abcb1a/1b, Abcg2, or both, respectively, and we studied plasma pharmacokinetics and brain accumulation. The systemic exposure on oral administration was not different among all strains. However, brain accumulation was 4.3-fold increased in Abcg2(-/-) mice and 9.3-fold increased in Abcb1a/1b;Abcg2(-/-) mice. Moreover, when wild-type mice were treated with sorafenib in combination with the dual P-gp and ABCG2 inhibitor elacridar, brain accumulation was similar to that observed for Abcb1a/1b;Abcg2(-/-) mice. These results show that the brain accumulation of sorafenib is primarily restricted by ABCG2. This contrasts with previous studies using shared ABCG2 and P-gp substrates, which all suggested that P-gp dominates at the blood-brain barrier, and that an effect of ABCG2 is only evident when both transporters are absent. Interestingly, for sorafenib, it is the other way around, that is, ABCG2, and not P-gp, plays the dominant role in restricting its brain accumulation. Clinically, our findings may be relevant for the treatment of renal cell carcinoma patients with central nervous system relapses, as a dual ABCG2 and P-gp inhibitor might improve the central nervous system entry and thereby the therapeutic efficacy of sorafenib.