Mathematical learning difficulties (MLD) are characterized by difficulties in the understanding and processing of numbers and quantities. While MLD is related mainly to numerical deficits, studies show that this population has several other cognitive difficulties. The current study examined whether the cognitive deficits consisting of cognitive instability in the form of intra-subject variability (ISV) will also characterize the performance of individuals with MLD.Female adults with MLD and a matched control group performed numerical and non-numerical tasks and various ISV measures were compared between the two groups.Overall, the results showed that participants with MLD had higher ISV measures, including SD, sigma, and tau, only when performing numerical tasks.It appears that the cognitive system of MLD participants is less consistent and noisier when performing these numerical tasks. However, this inconsistency is not a general deficiency but rather a numerically specific one, as this inconsistency does not seem to characterize the performance of individuals with MLD in tasks that do not involve numerical processing. These findings have unique importance for understanding the difficulties characterizing individuals with MLD and possible future interventions.