Although T cells can mediate potent antitumor responses, immune tolerance mechanisms often result in the deletion or inactivation of T cells that express T-cell receptors (TCRs) against potentially effective target epitopes. Various approaches have been devised to circumvent this problem. In one approach, the gene encoding an antibody against a cancer-associated antigen is linked, in the form of a single-chain variable fragment (scFv), to genes that encode transmembrane and signaling domains. This chimeric antigen receptor (CAR) is then introduced into T cells for adoptive T-cell therapy. In another approach, the anti-cancer scFv is fused to a scFv that binds to the CD3ε subunit of the TCR/CD3 complex. This fusion protein serves as a soluble, injectable product that has recently been termed bispecific T-cell engager (BiTE). Both strategies have now been tested in clinical trials with promising results, but the comparative efficacies are not known. Here, we performed a direct comparison of the in vitro sensitivity of each strategy, using the same anti-cancer scFv fragments, directed against a tumor-specific glycopeptide epitope on the sialomucin-like transmembrane glycoprotein OTS8, which results form a cancer-specific mutation of Cosmc. While both approaches showed specific responses to the epitope as revealed by T cell-mediated cytokine release and target cell lysis, CAR-targeted T cells were more sensitive than BiTE-targeted T cells to low numbers of antigens per cell. The sensitivity scale described here provides a guide to the potential use of these two different approaches.