This study examines the impact of genetically modified maize on labor, cost, and input substitutability for smallholders in South Africa. Data was collected during the 2009- 2010 maize production season from 184 households with a total of 212 maize plots in two regions, Hlabisa and Simdlangetsha, located in KwaZulu-Natal, South Africa. Producers of Roundup Ready® (RR) maize use significantly less child, female, and male labor than non- RR producers, resulting in lower costs in spite of significantly higher herbicide, seed, and fertilizer prices. An unrestricted cost function approach is used to evaluate the differences in cost between maize varieties, assuming that households use different input allocations to minimize cost while producing a fixed level of output. A treatment effects model used to control for selection bias shows that the entire cost advantage and more can be attributed to the Roundup Ready® technology. The treatment effects model reveals that RR maize producers have $102.44 (30%) lower costs per maize plot after taking into consideration the inverse Mills ratio, suggesting that the OLS model underestimated the cost-reducing effect of RR maize. Therefore, the entire cost advantage and more can be attributed to RR maize after isolating the effect of RR maize on total cost by disentangling the lower costs attributed to RR maize from those associated with farm and farmer characteristics. These results are confirmed using a nonparametric kernel density estimator. Elasticities of factor substitution indicate strong substitutability among inputs; however, lack of statistical significance limits interpretation of results.