The deployment of transgenic mosquitoes carrying genes for refractoriness to malaria has long been seen as a futuristic scenario riddled with technical difficulties. The integration of anti-malarial effector genes and a gene-drive system into the mosquito genome without affecting mosquito fitness is recognized as critical to the success of this malaria control strategy. Here we conducted detailed fitness studies of two Anopheles gambiae s.s. transgenic lines recently developed using a two-phase targeted genetic transformation system. In replicated cage-invasion experiments, males and females of the EE Phase-1 docking strain and EVida3 Phase-2 strain loaded with an antimicrobial peptide (AMP) expressed upon blood-feeding, were mixed with individuals of a recently-colonized strain of the Mopti chromosomal form. The experimental design enabled us to detect initial strain reproductive success differences, assortative mating and hybrid vigor that may characterize mosquito release situations. In addition, the potential fitness costs of the unloaded Phase-1 and loaded Phase-2 genetic constructs, independent of the strains' original genetic backgrounds, were estimated between the 1(st) instar larvae, pupae and adult stages over 10 generations. The Phase-1 unloaded docking cassette was found to have significantly lower allelic fitness relative to the wild type allele during larval development. However, overall genotypic fitness was comparable to the wild type allele across all stages leading to stable equilibrium in all replicates. In contrast, the Phase-2 construct expressing EVida3 disappeared from all replicates within 10 generations due to lower fitness of hemi- and homozygous larvae, suggesting costly background AMP expression and/or of the DsRed2 marker. This is the first study to effectively partition independent fitness stage-specific determinants in unloaded and loaded transgenic strains of a Phase-1-2 transformation system. Critically, the high fitness of the Phase-1 docking strain makes it the ideal model system for measuring the genetic load of novel candidate anti-malarial molecules in vivo.