This study, besides focusing on the applicability of computer-mediated Dynamic Assessment (DA), attempts to investigate the impacts of mediation provision in three different, albeit related, manners, that is, through computer, human, and mixture of both human and computer, on reading comprehension of L2 learners. In this mixed-methods study, a total of 20 students were purposively assigned to three groups. To collect the data, DIALANG as a placement test, two researcher-made software programs, and the researcher as a rater were used. For the qualitative part, the transcripts provided a rich analysis of the dialogic interactions between the mediator and the learners. As for the quantitative part, using analysis of covariance (ANCOVA), the posttest results were indicative of less mediation use and more non-mediated scores which were obtained through Non-Dynamic Assessment (NDA). Moreover, comparing the results of the human-only type (HO group) and computer-only type (CO group) revealed no statistical difference (p > 0.05), which is indicative of the fact that both types of mediations had a similar influence on the learners' reading comprehension. The results also indicated the effectiveness of the type of mediation which is offered through computers while a human mediator is present as well.