When it comes to creating convincing virtual environments, the implementation of advanced auralizations is an obvious goal. A recent study found that head-tracked binaural auralizations were rated close-to-real regarding acoustical properties such as reverberance, source distance, or overall quality (Blau et al., 2021). An open question is whether these auralizations also allow close-to-real sound-source localization in a virtual reality (VR) scenario and enhance social presence. Therefore, we implemented an eye-tracking paradigm in VR as a naturalistic and implicit tool to measure spatial attention and sound-source localization to investigate the impact of different auralizations. In this VR study, participants had to openly attend to a virtual agent in a seminar room to indicate the perceived sound-source localization while gaze direction was measured. We compared three highly plausible auralizations (Stärz et al., 2022) to loudspeakers and to an anchor (gaming audio engine). Besides the impact on accuracy of speech localization, the extent to which auralizations affect social presence was examined. The results can be seen as an impulse to endeavor sophisticated auralizations in the field of clinical psychology, since higher levels of presence supporting smoother social interactions may contribute to enhanced emotional processing in psychotherapeutic interventions using VR (Diemer et al., 2015).