Metabolic rates are typically thought to have important influences on fitness and more broadly be relevant to the ecology and evolution of animals. Previous studies demonstrate that metabolic rates are repeatable to a certain extent under constant conditions, but how social conditions influence the repeatability of metabolic rate remains largely unknown. In this study we investigated the repeatability of resting metabolic rate (RMR) in the highly-social crucian carp (Carassius auratus) after being socially separated for different time periods relative to control fish that were not socially separated. We found that RMR was repeatable in fish in the control group, while the repeatability of RMR disappeared quickly (even within 7 d) when fish were exposed to social separation. This study is the first to examine the role of social separation at different time periods on the repeatability of intra-individual physiological variation in fish. We highlight that inter-individual repeatability of metabolic rate can be substantial over time but was eliminated by social separation. The findings indicate that the repeatability of metabolic rate in fish is condition-dependent, and that the change in repeatability of metabolic rate should not be overlooked when considering the ecological and evolutionary effects of environmental change.