Background Quantitative magnetic resonance imaging (qMRI) of articular cartilage represents a powerful tool in osteoarthritis research, but has so far been confined to a field strength of 1.5 T. The aim of the study was to determine the reproducibility and accuracy of qMRI assessments of the knee cartilage volume by comparing quantitative swine cartilage volumes of the sagittal (sag) multi echo data imagine combination water-excitation (MEDICwe) sequence and the fast low-angle shoot water-excitation (FLASHwe) sequence at 3.0-T MRI to directly measured volumes (DMV) of the surgically removed articular cartilage.Methods Test-retest MRI was acquired in 20 swine knees. Two sag FLASHwe sequences and two sag MEDICwe sequences (spatial resolution 0.4 mm × 0.4 mm × 1.0 mm of 3-dimension (3D) were acquired at 3-T MRI in a knee.Articular cartilage volume was calculated from 3D reformations of the MRI by using a manual program. Calculated volumes were compared with DMV of the surgically removed articular cartilage. Knee joint cartilage plates were quantified paired in order.Results In the knee joint of swine, reproducibility errors (paired analysis) for cartilage volume were 2.5% to 3.2% with sag FLASHwe, and 1.6% to 3.0% with sag MEDICwe. Correlation coefficients between results obtained with qMRI and DMV ranged from 0.90 to 0.98 for cartilage volume. Systematic pairwise difference between results obtained with qMRI and DMV ranged from -1.1% to 2.8%. Random pairwise differences between results obtained with qMRI and DMV ranged from (2.9 ±2.4)% to (6.8±4.5)%.Conclusions FLASHwe and MEDICwe sequences permit highly accurate and reproducible analysis of cartilage volume in the knee joints of swine at 3-T MRI. Cartilage volume reproducibility for the MEDICwe data is slightly higher than the FLASHwe data.