OBJECTIVE: To investigate the dose-response relationship of semaglutide versus placebo and open-label liraglutide in terms of glycemic control in patients with type 2 diabetes. RESEARCH DESIGN AND METHODS: This was a 12-week, randomized, double-blind phase 2 trial. Patients (n = 415) were randomized to receive a subcutaneous injection of semaglutide once weekly without dose escalation (0.1-0.8 mg) or with dose escalation (E) (0.4 mg steps to 0.8 or 1.6 mg E over 1-2 weeks), open-label liraglutide once daily (1.2 or 1.8 mg), or placebo. The primary end point was change in HbA1c level from baseline. Secondary end points included change in body weight, safety, and tolerability. RESULTS: Semaglutide dose-dependently reduced the level of HbA1c from baseline (8.1 ± 0.8%) to week 12 by up to -1.7%, and body weight by up to -4.8 kg (1.6 mg E, P < 0.001 vs. placebo). Up to 81% of patients achieved an HbA1c level of <7%. HbA1c level and weight reductions with semaglutide 1.6 mg E were greater than those with liraglutide 1.2 and 1.8 mg (based on unadjusted CIs), but adverse events (AEs) and withdrawals occurred more frequently. The incidence of nausea, vomiting, and withdrawal due to gastrointestinal AEs increased with the semaglutide dose; most events were mild to moderate, transient, and ameliorated by dose escalation. There were no major episodes of hypoglycemia and few cases of injection site reactions. CONCLUSIONS: After 12 weeks, semaglutide dose-dependently reduced HbA1c level and weight in patients with type 2 diabetes. No unexpected safety or tolerability concerns were identified; gastrointestinal AEs typical of glucagon-like peptide 1 receptor agonists were mitigated by dose escalation. On this basis, weekly semaglutide doses of 0.5 and 1.0 mg with a 4-week dose escalation were selected for phase 3.