The aim of the study was to describe variations in swept visual evoked potential (SWEEP-VEP) assessment of visual acuity and contrast sensitivity in infants and to evaluate the best way to estimate visual performance from obtained SWEEP-VEP data. The visual performance of 92 infants (6-40 wk of age) was measured in two separate visits. Results were verified with repeated tests in seven adults. There was a strong association between the two measurements of infant visual acuity (r = 0.91, p < 0.001), with no constant bias and an inter-assay coefficient of variation of 8.4%. The intra-assay coefficient of variation was 17% and in repeated sessions all obtained acuity measures were normally distributed, indicating that the mean and not the maximum threshold best estimates visual acuity. This estimate of visual acuity also had lower test-retest variability than those calculated from the maximum threshold or threshold from the average EEG signals (p = 0.001). Test-retest measures of infant contrast sensitivity had a correlation coefficient of 0.72 (p < 0.001) and an inter-assay coefficient of variation of 23%. With the observed test-retest variability, SWEEP-VEP is less valid for estimating the visual performance of individual subjects, but it can give reliable group means. This method was well suited to describe visual development in the infants, which for acuity as well as contrast sensitivity increased by 0.64 octave per doubling in age. However, the variability of the SWEEP-VEP method can be a limiting factor, for example, in the assessment of the potential effect of dietary docosahexaenoic acid in a homogeneous group of infants.