Tumour immunotherapy, and particularly immue checkpoint inhibitors, have resulted in considerable response rates in patients with metastatic cancer. However, most of these approaches are limited to immunogenic tumours. Based on its ability to stimulate cytotoxic T cells, interleukin-2 (IL-2) has been used to treat patients with metastatic melanoma and metastatic kidney cancer. Clinical efficacy achieved through high doses is countered by severe adverse effects on vascular endothelial cells and various organs, a short in vivo half-life, and the stimulation of regulatory T cells that counteract antitumour immune responses. Accumulating evidence suggests that IL-2 receptor β (CD122)-biased IL-2 formulations address the shortcomings of IL-2 cancer immunotherapy. This knowledge stems from studies using CD122-biased IL-2/anti-IL-2 antibody complexes (IL-2 complexes), which preferentially stimulate CD8+ T cells, while interaction with regulatory T cells and vascular endothelial cells is disfavoured by the anti-IL-2 antibody used. CD122-biased IL-2 complexes, when assessed in different mouse cancer models, cause stronger antitumour effects and significantly less adverse effects than high-dose IL-2. A recently developed and characterised anti-human IL-2 antibody, termed NARA1, forms human CD122-biased IL-2 complexes. Alternative strategies based on this concept, such as site-directed pegylation and mutation of IL-2, have also been pursued. Moreover, recent data have shown that a combination of CD122-biased IL-2 formulations with immune checkpoint inhibitors, antigen-specific immunotherapy and epigenetic modifying drugs results in synergistic anti-cancer effects in various tumour models. Thus, CD122-biased IL-2 approaches constitute a novel class of immunotherapy for metastatic cancer that has the potential to complement and increase the efficacy of other antitumour strategies.