Cerebral laterality, via hemispheric specialisation, has been evidenced across the animal kingdom and linked to cognitive performance in a number of species. Previously it has been suggested that cognitive processing is more efficient in brains with stronger hemispheric differences in processing, which may be the key fitness benefit driving the evolution of laterality. However, evidence supporting a positive association between cognitive performance and lateralization is mixed: data from studies of fish and birds show a positive relationship whereas more limited data from studies of mammals suggest a weak or even negative relationship, suggesting the intriguing possibility of a mammal/non-mammal divide in the nature of this relationship. Here, we report an empirical test examining the relationship between lateralization and cognitive performance in wild grey squirrels ( Sciurus carolinensis ) by measuring left/right paw preference as a behavioural assay of cerebral lateralization and learning speed as an assay of cognitive efficiency. We carried out a motor-based laterality test using a reaching paradigm and measured learning speed on a problem-solving task. In accordance with the suggestion of a mammal/non-mammal divide, we found a negative relationship between strength of paw preference and performance on the learning task. We discuss this finding in light of niche-specific adaptations, task-specific demands and cognitive flexibility.