The ability of IL-12 to initiate anti-leukaemia immune responses has been well established; however clinical outcomes fail to recapitulate the therapeutic benefits observed in the laboratory. To address this, we compared two systems of IL-12 therapy that elicit protective immune responses against the murine acute lymphoblastic leukaemia (ALL) cell line, 70Z/3. These systems differ in the method of IL-12 administration and ultimately result in leukaemia clearance by distinct mechanisms, emphasizing the importance of treatment vehicle. Injecting low-dose IL-12 was sufficient to elicit long-term protective immunity against an established leukaemia burden, mediated by both CD4(+) and CD8(+) T cells. These findings agree with the standard model of IL-12 activity. We compared this protocol to a cell-based approach in which a novel lentiviral vector (LV) expressing murine IL-12 was created, 70Z/3 cells transduced, and clones selected that stably secrete different amounts of IL-12. We found that only a small proportion (1%) of IL-12 secreting cells were required for rejection but that the amount of IL-12 produced per cell was critical for successful therapy. Importantly, the levels of IL-12 required were found to be higher than the levels reported to date in the human clinical trial literature. We found that the cell-based approach led to protective immunity that was both long-term and specific but dependent primarily on a CD4(+) cellular subset alone. Our results highlight that the mode of IL-12 delivery has a distinct impact on the immune response initiated, leading to leukaemia clearance by disparate mechanisms. We also establish a new and critical parameter, IL-12 production/cell, which may have significant implications for future therapeutic design.