Traditional RNA-DNA chimeric oligonucleotides (chimeraplasts), composed of a continuous stretch of RNA and DNA residues in a duplex conformation, have been shown to correct single-base mutations in episomal and genomic DNA both in vitro and in vivo. In the current study, we have compared the efficiency of single-base pair correction between a traditionally designed chimeraplast (covalently linked duplex) and hybrid chimeraplasts (noncovalent duplexes formed from stretches of RNA and DNA nucleotides synthesized individually and hybridized in vitro). Six hybrid chimeraplasts of identical length were constructed with various lengths of target homology and strand location of the desired nucleotide change. These constructs were evaluated for their ability to correct a point mutation in the gene encoding recombinant enhanced green fluorescent protein (eGFP) that rendered the protein nonfluorescent. A plasmid encoding this mutant eGFP gene and a chimeraplast were co-introduced directly into the nuclei of primary fibroblasts by microinjection. As shown by the recovery of eGFP fluorescence, three of the six hybrid chimeraplasts demonstrated the ability to mediate gene correction (0.4–2.4%). Covalent joining of RNA and DNA strands in chimeraplasts was not necessary for correction of DNA mutations. However, the strand placement of the desired nucleotide change and the length of nonhomologous sequences flanking target nucleotides played a crucial role in the efficiency of chimeraplast-mediated gene correction. Despite the ability of certain chimeraplast designs to correct point mutations in episomal plasmids, targeted correction of integrated copies of the mutant eGFP transgene was unsuccessful in primary fibroblasts. These results demonstrate that, although chimeraplasts are fairly effective at targeting episomal DNA in primary cells, further optimization is required to increase the efficiency for targeting integrated genes.