Both noncontrast and contrast-enhanced approaches to gated thoracic magnetic resonance angiography (MRA) for aortic root evaluation have been reported at 3T. We compare qualitative and quantitative image quality measures for the two approaches, and assess the reproducibility of standard aortic measurements. Respiratory and cardiac gated MRA of the chest was performed at 3T in 45 patients: 23 after administration of iron-based blood pool contrast, and 22 without contrast. Image quality was assessed with a 5-point Likert scale, vessel lumen-to-muscle contrast ratios, and vessel wall sharpness. Two reviewers measured the ascending aorta diameter and valve annulus area. Interrater agreement was assessed using Bland–Altman plots and coefficient of variation (CV). Qualitative image quality was better with blood pool contrast in all principal vessels of the chest (mean Likert of 4.20 ± 0.79 vs. 2.60 ± 0.77, p < 0.001). Quantitative assessment was also improved with higher contrast ratios in all vessels (5.26 ± 3.3 vs. 1.90 ± 0.53, p < 0.001), and greater sharpness of the aortic annulus and ascending aorta (0.70 ± 0.16 vs. 0.56 ± 0.14 mm−1, p < 0.001, and 0.87 ± 0.16 vs. 0.62 ± 0.16 mm−1, p = 0.008, respectively). Reproducibility of measurement was marginally better for the ascending aorta diameter (CV of 2.80 vs. 3.23%), but substantially increased for the aortic valve annulus area with blood pool contrast (CV of 4.93 vs. 7.32%). The use of a blood pool contrast agent for gated thoracic MRA improves image quality compared to a noncontrast technique, and provides more reproducible measurements of the aortic valve annulus area.