Nitrogen (N) balance experiment was carried out to study the individual and group response of growing pigs to lysine (Lys) intake. A series of 15 purified diets limiting in Lys were fed sequentially to nine pigs (initial body weight, 39 kg) during a 15-day experiment. Lysine concentration ranged from 50% to 140% of its assumed requirement. N retention in individual pigs was related to Lys intake using linear-plateau (LP) and quadratic-plateau (QP) models. No difference in the closeness of fit between the models was found (p = 0.72). There were significant differences between the slopes of individual regression lines in the LP model (p = 0.018) and between the individual plateau values in both models (p < 0.0001). The breakpoint values in the QP model were greater than those of the LP model (p = 0.027), while the plateau values estimated by both models were similar (p = 0.32). In the LP model, no relationship was found between the slopes and plateau values (p = 0.67, r = 0.17) or between the slopes and breakpoint values (p = 0.55, r = -0.23), thus suggesting that pigs with higher protein deposition rate do not utilize lysine more efficiently. However, there was a close positive correlation between breakpoint and plateau values (p = 0.0005, r = 0.92). Based on LP and QP models, the mean daily requirement of Lys for a 47-kg gilt was estimated to be 20.8 and 24.2 g (0.79 and 0.92 g/MJ metabolizable energy), respectively, with coefficients of variation of approximately 10%. Marginal efficiency of Lys utilization derived from the LP model was 0.67 and was dependent on the range of input data selected. Lysine disappearance was a curvilinear function of Lys intake, indicating that Lys catabolism is not directly related to Lys intake. The closeness of fit of exponential, saturation kinetics or four-parameter logistic models applied to data set of all pigs was similar to that of LP or QP models.