This study explored the effectiveness of commercial software in teaching grammar as compared to blended and teacher-led learning conditions, and the attitudes of students towards using commercial software to learn grammar. The study was conducted with a participant teacher and 42 upper-intermediate level preparatory school students at Yıldız Technical University School of Foreign Languages, who were assigned to one of the instruction groups, which were computerbased, teacher-led and blended. A three-week procedure of grammar teaching was carried out according to the groups of the participants through materials developed by the researcher. The data were gathered via a pre-test, three immediate post-tests, a delayed posttest and a students’ attitude questionnaire. Following the pre-test, the computer-based group was given only computer-based instruction. This group reviewed and practiced the target structures through the commercial software. The teacher-led group was given instruction by the participant teacher. They reviewed and practiced the target structures with the teacher in the classroom. The blended group was given instruction via the participant teacher. They reviewed and practiced the target structures through the commercial software. All the participants were given immediate post-tests right after the procedure. The delayed post-test was administered two weeks after the procedure ended. One week later, they were administered the attitude questionnaire. The results of the quantitative analysis revealed that the teacher-led instruction was slightly more effective than the computer-based and blended learning conditions. The results also indicated that the students’ attitudes towards using commercial software to learn grammar were negative. This study implied that further research is needed to integrate computer-assisted language instruction into our educational systems in different ways after eliminating its disadvantages, which may negatively affect students’ attitudes.