Human epidermal growth factor receptor type 3 (HER3) is recognized to be involved in resistance to HER targeting therapies. A number of HER3-targeting monoclonal antibodies are under clinical investigation as potential cancer therapeutics. Smaller high-affinity scaffold proteins are attractive non-Fc containing alternatives to antibodies. A previous study indicated that anti-HER3 affibody molecules could delay the growth of xenografted HER3-positive tumors. Here, we designed a second-generation HER3-targeting construct (TAM-HER3), containing two HER3-specific affibody molecules bridged by an albumin-binding domain (ABD) for extension of blood circulation. Receptor blocking activity was demonstrated in vitro. In mice bearing BxPC-3 xenografts, the therapeutic efficacy of TAM-HER3 was compared to the HER3-specific monoclonal antibody seribantumab (MM-121). TAM-HER3 inhibited heregulin-induced phosphorylation in a panel of HER3-expressing cancer cells and was found to be equally as potent as seribantumab in terms of therapeutic efficacy in vivo and with a similar safety profile. Median survival times were 60 days for TAM-HER3, 54 days for seribantumab, and 41 days for the control group. No pathological changes were observed in cytopathological examination. The multimeric HER3-binding affibody molecule in fusion to ABD seems promising for further evaluation as candidate therapeutics for treatment of HER3-overexpressing tumors.