Background: The ARGO study was a phase II, double-blind, placebo controlled, four parallel arm trial of tideglusib in Alzheimer's disease (AD). Objective: To prove the clinical efficacy of an inhibitor of glycogen synthase kinase-3 (GSK-3), in AD. Methods: Mild to moderate (Mini-Mental State Examination (MMSE) score, 14-26) AD patients on cholinesterase inhibitor and/or memantine treatment were administered tideglusib or placebo for 26 weeks. The ADAS-cog15 was the primary efficacy measure; function, cognition, behavior, and quality of life were assessed as secondary measures; cerebral atrophy in MRI and the levels of tau, amyloid-β, and BACE1 in cerebrospinal fluid (CSF) were exploratory endpoints. Results: 306 AD patients were randomized to active (1000 mg QD: n = 86, 1000 mg QOD: n = 90, and 500 mg QD: n = 50) or placebo (n = 85) in 55 sites in four European countries. There were no statistically significant differences between either active and placebo arms in the efficacy variables. However, BACE1 in CSF significantly decreased with treatment in a small subgroup of patients. Participants with mild AD in the 500 mg QD group showed significant responses on ADAS-cog15, MMSE, and word fluency. Diarrhea (14-18% in active, 11% placebo) and dose-dependent, mild to moderate, and fully reversible transaminase increase (9-16% in active, 3.5% placebo) were the most frequent adverse events. Conclusions: Short term (26 weeks) tideglusib was acceptably safe but produced no clinical benefit in this trial. However, given the non-linear dose response, especially in mildly affected patients, further dose finding studies in early disease stages and for longer duration are warranted to examine GSK-3 inhibition in AD patients.