While engagement of the inhibitory Fcγ-receptor (FcγR) IIB is an absolute requirement for in vivo antitumor activity of agonistic mouse anti-CD40 monoclonal antibodies (mAbs), a similar requirement for human mAbs has been disputed. By using a mouse model humanized for its FcγRs and CD40, we revealed that FcγRIIB engagement is essential for the activity of human CD40 mAbs, while engagement of the activating FcγRIIA inhibits this activity. By engineering Fc variants with selective enhanced binding to FcγRIIB, but not to FcγRIIA, significantly improved antitumor immunity was observed. These findings highlight the necessity of optimizing the Fc domain for this class of therapeutic antibodies by using appropriate preclinical models that accurately reflect the unique affinities and cellular expression of human FcγR. Display omitted •Wild-type mice fail to model the activity of human CD40 IgG subclasses•Humanized FcγR/CD40 mice enabled assessment of hFcR contribution to CD40 mAb activity•Binding affinity to FcγRIIB dictates the agonistic activity of human CD40 IgGs•CD40 mAbs enhanced for FcγRIIB, but not FcγRIIA, have optimal antitumor activity Dahan et al. generate a mouse model fully humanized for FcγRs and CD40. Using this model, they discover that FcγRIIB engagement is essential for the antitumor activity of human CD40 agonistic monoclonal antibodies (mAbs), whereas FcγRIIA engagement is inhibitory. They use this knowledge to improve these mAbs.