This study compared five non-linear growth functions (Gompertz, logistic, Lopez, Richards and Weibull) using body weight (BW) measurements from a commercial strain (Ross 308) and Iranian native chickens. Seventy two commercial broilers and seventy two native chickens were randomly assigned to four treatments. Each treatment consisted of six replicates of three chickens. The chickens were fed with maize-soyabean meal supplying four levels of metabolizable energy (ME) (100%, 94%, 88% and 82% of NRC recommendations) for eight weeks. Body weight was measured weekly from 2 weeks of age. Flexible growth functions were evaluated for their ability to describe the relationship between liveweight and age, and were compared with functions having a fixed point of inflection. Based on goodness of fit criteria and statistical performance, the flexible growth functions fit the data better than the functions with a fixed point of inflection, such as the Gompertz and logistic functions. Among the flexible growth functions, the Richards function gave the best fit to the data, with adjusted coefficients of determination ranging from 99.51 to 99.12 for commercial and native birds, respectively. The results indicated that commercial broilers had a higher final body weight (W_f) than the native chickens. The chickens receiving 100% ME concentrations had higher body weights compared with the chickens on lower energy levels. Based on the Richards function, the decreased dietary ME concentrations caused a linear reduction in W_f in the commercial strain (P < 0.05); however, dietary energy concentrations had no effect on the W_f and other growth curve parameters in native chickens (P > 0.05).