The artificial pancreas (AP) in its 2 versions, single-hormone AP (insulin only; SAP) and dual-hormone AP (insulin and glucagon; DAP), is a promising modality for the treatment of type 1 diabetes (T1D). We conducted an open-label, randomized, crossover study comparing SAP and DAP in 17 adults (age of 37.2±13.6 years, HBA1c 8.0±1.0%, during 2 types of exercises reported to have different hypoglycemic risks: moderate-intensity continuous exercise (60% VO2peak for 60 min) and high-intensity interval exercise (2 min alternating intervals at 85% and 50% VO2peak for 40 min, with 2×10 min at 45% VO2peak at the start and the end of the sessions). SAP and DAP were applied from 15:30 till 19:30, exercise started at 18:00 and announced 20 minutes earlier to the AP systems. During SAP as compared to DAP: exercise-induced hypoglycemia (plasma glucose ≤3.3 mmol/L with symptoms or <3 regardless of symptoms) was observed in 31.25% (10/32) vs. 9% (3/33) of interventions (p=0.02); the median (IQR) percentage of time with glucose <4.0 mmol/L was 11 (0 to 46.7)% vs. 0 (0 to 0)% (p=0.0001); time with glucose between 4 and 10 mmol/L was 71.4 (53.2 to 100)% vs. 100 (100 to 100)% (p=0.003); the median area under the curve for glucose <4.0 mmol/L was 31.6 (0 to 153.9 vs. 0 (0 to 0) mmol/L×min per h (p=0.0001). Higher doses of glucagon were needed during the continuous exercise 14.4 (9.5 to 16.9) mg vs. the interval exercise sessions 8.5 (4.6 to 16.9) mg (p=0.03). In summary, DAP is offering a tighter control and a better potential to prevent hypoglycemia during exercise in adults with T1D.