Background: The Medical Model of disability focuses on diagnosed conditions. It is used in policy particularly to categorise people. This enables predictions and forecasting about the size of policy needs but tends to homogenise disability representations, assigning a negative evaluation to illness that may be irrespective of patho-anatomical correlates. The Social Model considers disability as imposed by society through attitudes and barriers. The Neurodiversity Model is a type of social and cultural model with biological implications; it states that differences in brain and behaviour lie on a non-pathological spectrum. Critics say this whitewashes lived experience. Policymakers may devalue the Neurodiversity Model's origins within activist neurodiverse communities. The model that policy and practice decision makers use has fundamental effects on their impacts. Aims and objectives: The Medical and Neurodiversity Models are reviewed in reference to their politicisation as ways to characterise disability, and identity politics. The implications socially and for disability policy and practice and evidence use are considered. Key conclusions: Both models fall short in addressing the needs of the broad community of the disabled, yet both have useful features. We propose the Biological Gradient Model (BGM), which integrates scientific theory while avoiding pathology-based concepts and value-laden judgments concerning 'deficiency'. Its usefulness is demonstrated; it resolves some of the ambiguity and tensions that exist in the way people with disability are viewed by different groups and treated within policy. It has the potential to reduce issues of partial representation, where the voices of those who cannot self-advocate may be less heard.