Although metformin is widely used for treatment of type 2 diabetes (T2D), its glucose-lowering mechanism remains unclear. Using the glucagon-like peptide 1 (GLP-1) receptor (GLP-1R) antagonist exendin(9-39)NH2, we tested the hypothesis that postprandial GLP-1-mediated effects contribute to the glucose-lowering potential of metformin in T2D. In a randomized, placebo-controlled, double-blind, crossover study, 15 individuals with T2D (median HbA1c 50 mmol/mol [6.7%], body mass index 30.1 kg/m2, age 71 years) underwent, in randomized order, 14 days of metformin and placebo treatment, respectively. Each treatment period was preceded by 14 days without any glucose-lowering medicine and concluded by two 4 h mixed meal tests performed in randomized order and separated by >24 h with either continuous intravenous exendin(9-39)NH2 or saline infusion. Compared to placebo, metformin treatment lowered fasting plasma glucose (mean of differences [MD] 1.4 mmol/L × min [95% CI 0.8-2.0]) as well as postprandial plasma glucose excursions during both saline infusion (MD 186 mmol/L × min [95% CI 64-307]) and exendin(9-39)NH2 infusion (MD 268 mmol/L × min [95% CI 108-427]). The metformin-induced improvement in postprandial glucose tolerance was unaffected by GLP-1R antagonization (MD 82 mmol/L × min [95% CI -6564-170]). Metformin treatment increased fasting plasma GLP-1 (MD 1.7 pmol/L × min [95% CI 0.39-2.9]) but did not affect postprandial GLP-1 responses (MD 820 pmol/L × min [95% CI -1750-111]). Using GLP-1R antagonization, we could not detect GLP-1-mediated postprandial glucose-lowering effect of metformin in individuals with T2D. We show that 2 weeks of metformin treatment increases fasting plasma GLP-1, which may contribute to metformin's beneficial effect on fasting plasma glucose in T2D. Trial registration: Clinicaltrials.gov NCT03246451.