Abstract Prostate cancer is considered immunologically “cold”, with low mutational burden, tumor-infiltrating immune cells, and PD-L1 levels, culminating in poor response to immune checkpoint therapies. Bispecific CD3 redirection antibodies can elicit T cell-mediated cytotoxicity and hold promise for immune cell recruitment into prostate tumors. CD3 redirection antibodies in solid tumors are still in early phases of clinical development, and it is not yet understood whether these potential therapies will achieve the high response rates observed in hematological malignancies or result in durable T cell responses. Here we demonstrated that treatment with a PSMA targeted CD3 redirector resulted in efficacy against LnCaP.AR human prostate xenografts in CD34+ cord blood humanized mice. Efficacy correlated with T cell infiltration into tumors with an activated phenotype, but also increased PD-L1 expression. Engineered overexpression of PD-L1 in LNCaP.AR tumors resulted in resistance to PSMAxCD3 bispecific antibody treatment, whereas sensitivity was restored in combination with anti-PD-1 antibody pembrolizumab. PSMAxCD3 and anti-PD-1 combination treatment resulted in complete tumor responses in approximately 20% of mice, and elicited immune responses that delayed growth of rechallenged tumors. In a second prostate model, patient derived LuCaP 86.2 xenografts, PSMAxCD3 monotherapy treatment resulted in complete responses in 25% of mice. When PSMAxCD3-treated responder mice were rechallenged with LuCaP 86.2 tumors, partial control of tumor regrowth was associated with expansion of effector memory T cells. These studies show that PSMAxCD3 treatment elicits antitumor memory T cell responses, and that combination with PD-1 blockade can enhance these effects in tumors with immune suppressive tumor microenvironments.