Objective: To compare the outcomes and costs of clustered domestic and standard Australian models of residential aged care. Design: Cross‐sectional retrospective analysis of linked health service data, January 2015 – February 2016. Setting: 17 aged care facilities in four Australian states providing clustered (four) or standard Australian (13) models of residential aged care. Participants: People with or without cognitive impairment residing in a residential aged care facility (RACF) for at least 12 months, not in palliative care, with a family member willing to participate on their behalf if required. 901 residents were eligible; 541 consented to participation (24% self‐consent, 76% proxy consent). Main outcome measures: Quality of life (measured with EQ‐5D‐5L); medical service use; health and residential care costs. Results: After adjusting for patient‐ and facility‐level factors, individuals residing in clustered models of care had better quality of life (adjusted mean EQ‐5D‐5L score difference, 0.107; 95% CI, 0.028–0.186; P = 0.008), lower hospitalisation rates (adjusted rate ratio, 0.32; 95% CI, 0.13–0.79; P = 0.010), and lower emergency department presentation rates (adjusted rate ratio, 0.27; 95% CI, 0.14–0.53; P < 0.001) than residents of standard care facilities. Unadjusted facility running costs were similar for the two models, but, after adjusting for resident‐ and facility‐related factors, it was estimated that overall there is a saving of $12 962 (2016 values; 95% CI, $11 092–14 831) per person per year in residential care costs. Conclusions: Clustered domestic models of residential care are associated with better quality of life and fewer hospitalisations for residents, without increasing whole of system costs.