Animal models are indispensable tools for Alzheimer disease (AD) research. Over the course of more than two decades, an increasing number of complementary rodent models has been generated. These models have facilitated testing hypotheses about the aetiology and progression of AD, dissecting the associated pathomechanisms and validating therapeutic interventions, thereby providing guidance for the design of human clinical trials. However, the lack of success in translating rodent data into therapeutic outcomes may challenge the validity of the current models. This Review critically evaluates the genetic and non-genetic strategies used in AD modelling, discussing their strengths and limitations, as well as new opportunities for the development of better models for the disease.