Cognitive deficits are common in patients with bipolar disorder (BD) in remission and may be associated with glycogen synthase kinase-3 (GSK-3) activity, which is inhibited by lithium. GSK-3 may be a relevant treatment target for interventions tailored at cognitive disturbances in BD but the relation between GSK-3 activity, cognition and lithium treatment is unknown. We therefore investigated the possible association between GSK-3 activity and cognition and whether lithium treatment moderates this association in patients with BD. In a prospective 6-12 month follow-up study, GSK- 3β activity in peripheral blood mononuclear cells was measured concurrently with cognitive performance assessed using a comprehensive test battery in 27 patients with BD-I in early and late remission following a manic or mixed episode. The GSK-3β activity, measured as serine-9 phosphorylated GSK-3β (pGSK-3β) and the GSK-3β ratio (serine-9-pGSK-3β /total GSK-3β), was negatively associated with sustained attention (p = 0.009 and p = 0.042, respectively), but not with other cognitive domains or global cognition. A crossover interaction between lithium treatment and the GSK activity was observed, indicating that lower pGSK-3β levels (p = 0.015) and GSK ratio (p = 0.010) were associated with better global cognition in lithium users whereas the opposite association was observed in non-lithium treated patients. Findings were not statistically significant after Bonferroni correction. In conclusion, cognitive functioning may be associated with GSK-3 activity in patients with BD-I and lithium treatment may modulate this relationship. Future studies in larger sample sizes are warranted to confirm these associations.