The objectives were to study the behavior of fifteen pre-commercial upland maize hybrids, analyze their agronomic performance regarding grain yield, and evaluate productivity components, as well as morpho-agronomic characteristics, in the Midwest Region of Brazil. Two experiments were conducted in the municipalities of Formosa-GO and Planaltina-DF, 2016/17 crop year. Both consisted of five pre-commercial maize hybrid platforms (HPA252, HPB262, HPB621, HPB646, and HPD354). Each platform consisted of three different versions: conventional, transgenic with a Bt gene that expresses the protein Cry1F, and transgenic with two Bt genes that express the proteins Cry1F and Cry1AB. The experiment was randomized blocks with four replications. The experimental plot was four lines five meters long considering the two central lines as useful. The lines were spaced 0.75 meters apart, and the final density was five plants per linear meter. To estimate grain yield, the plots were harvested, and the weight was extrapolated to kg.ha-1. The moisture was standardized at 14%. Data were submitted to analysis of variance, and the means were compared by Tukey test at 5% probability using the Sisvar software. Grain yield between treatments ranged from 8,381 to 12,908 kg.ha-1, and the average yield was 11,234 kg.ha-1. The parameters evaluated were divided into two groups. The first group contained parameters determining grain yield: grain depth, thousand grains weight, number of rows of grains, number of grains per row, and grain yield. The second group contained morpho-agronomic parameters that directly interfere with resistance to lodging and plant breaking: plant height, ear insertion height, and mean stem diameter. There was no direct effect of the transgenes on the evaluated hybrids since the classification of productivity was not divided into conventional and transgenic classes. It is suggested that the HPA252YH, HP621H and HP646H versions be discarded because grain yield performance was unsatisfactory compared to their respective conventional and transgenic counterparts. The parameters GD and W1000 are more effective for grain yield estimation than NR and NG. It was observed for HP621H that, in addition to a lower grain yield, there was a significant reduction in stem diameter, indicating a possible reduction in lodging tolerance and/or plant breaking when exposed to adverse climatic conditions such as windstorms. For the HPD354H version, the release of the transgenic counterpart HPD354YH is preferentially indicated since the H version had a significant reduction in stem diameter and a possible greater tolerance to lodging and/or plant breaking under the same conditions.