Modelling of human diseases is an essential component of biomedical research, to understand their pathogenesis and ultimately, develop therapeutic approaches. Here, we will describe models of tumours of the central nervous system, with focus on intrinsic CNS tumours. Model systems for brain tumours were established as early as the 1920s, using chemical carcinogenesis, and a systematic analysis of different carcinogens, with a more refined histological analysis followed in the 1950s and 1960s. Alternative approaches at the time used retroviral carcinogenesis, allowing a more topical, organ‐centred delivery. Most of the neoplasms arising from this approach were high‐grade gliomas. Whilst these experimental approaches did not directly demonstrate a cell of origin, the localisation and growth pattern of the tumours already pointed to an origin in the neurogenic zones of the brain. In the 1980s, expression of oncogenes in transgenic models allowed a more targeted approach by expressing the transgene under tissue‐specific promoters, whilst the constitutive inactivation of tumour suppressor genes (‘knock out’)‐often resulted in embryonic lethality. This limitation was elegantly solved by engineering the Cre‐lox system, allowing for a promoter‐specific, and often also time‐controlled gene inactivation. More recently, the use of the CRISPR Cas9 technology has significantly increased experimental flexibility of gene expression or gene inactivation and thus added increased value of rodent models for the study of pathogenesis and establishing preclinical models.