Self-efficacy represents an individual's belief that he or she can perform a particular task in a given domain. It is a strong predictor for performance and persistence in STEM education. Research shows that there is a large and persistent gender gap in student self-efficacy in STEM academic disciplines. In some cases, active teaching strategies have been shown to positively affect physics self-efficacy but it is unclear how this impact differs between men and women. In this study, we investigate the impact of a physics class taught with active teaching strategies on students' self-reported physics self-efficacy and how this impact varies across gender lines. We measured the change in physics self-efficacy over four different dimensions; conceptual understanding, problem-solving, lab and hands-on activities, and collaborative work. We report three main findings. First, the initial physics self-efficacy gender gap disappeared by the end of the semester. Second, female students' self-efficacy improved significantly, whereas there was no significant change in the self-efficacy of the male students. Third, the gender gaps in the conceptual understanding and problem-solving dimensions in particular, were significantly reduced. This study represents an initial step towards understanding the influence of active teaching strategies can have on reducing the self-efficacy gender gap.