Over the past half-century, modelling has come to play an increasingly important role in cancer research. These representational tools frame perceptions of malignant disease, guide public health responses, and help determine which interventions are necessary. But what makes a cancer model a model? What authority do they have? What stories do they tell? And how do they shape our understanding of disease and bodies? To shed light on these questions, this article explores the long history of cancer modelling in sub-Saharan Africa: a place where malignant disease has often been imagined as different, and where experimentation and improvisation in cancer research and treatment has been rife. Drawing on archival and ethnographic sources, we examine modelling strategies that health professionals have used to generate information about cancer in Africa from the mid 20th century to the present day. Focusing on three different case studies – anatomical models of Burkitt's lymphoma patients, diagnostic models for Kaposi sarcoma, and statistical models of the African smoking and lung cancer epidemic – we meditate on the multiplicity of models and modelling by identifying the epistemic strands that hold these representations together, as well as what sets them apart. In addition to contributing to discussions of how cancer research has taken shape beyond the Anglo-American realm, our article helps expand and complicate our understandings of what a disease model is.