Individual rare diseases may affect only a few people, making them difficult to recognize, diagnose or treat by studying humans alone. Instead, model organisms help to validate genetic associations, understand functional pathways and develop therapeutic interventions for rare diseases. In this Editorial, we point to the key parameters in face, construct, predictive and target validity for accurate disease modelling, with special emphasis on rare disease models. Raising the experimental standards for disease models will enhance successful clinical translation and benefit rare disease research.