Background Autism spectrum disorders affect more than one percent of the population, impairing social communication and increasing stereotyped behaviours. A micro-deletion of the 16p11.2 BP4-BP5 chromosomic region has been identified in one percent of patients also displaying intellectual disabilities. In mouse models generated to understand the mechanisms of this deletion, learning and memory deficits were pervasive in most genetic backgrounds, while social communication deficits were only detected in some models. Based on previous study (Arbogast et al. 2016 PLoS genetics), we selected the mouse model of 16p11.2 deletion on a hybrid C57BL/6NxC3B genetic background to itemize the social deficits. We examined whether behavioural deficits observed in short observation periods were representative of the phenotype displayed by the same mice over long-term monitoring. We recorded the individual and social behaviours of 16p11.2 Del/+ mice and their wild-type littermates from both sexes in short-term (15 min) and long-term (over two and three consecutive nights) social interactions of familiar mixed-genotype quartets of males and of females, and of same-genotype unfamiliar female pairs. Results We observed that Del/+ mice of both sexes increased significantly their activity compared to wild-type littermates only over long-term monitoring. In the social domain, Del/+ mice of both sexes displayed only limited impairments over short-term monitoring, and more visible deficits over long-term monitoring. When recorded in quartets of familiar individuals, social impairments were stronger in males than in females. In pairs, significant perturbations of the organisation of the social communication and behaviours in Del/+ females appeared mostly over the long-term. Conclusions Altogether, this suggests that social and contextual variations affect the phenotype of the 16p11.2 Del/+ mice differently in the activity and the social domains. The social behaviour was also differently affected between the two sexes. These findings confirm the importance of testing models both in short- and long-term conditions to provide a comprehensive view of their phenotype that will be more robust for pre-clinical targeted therapeutic trials.