The inconsistent findings about learning outcomes in VR (virtual reality) learning necessitate further robustness of empirical data. This article addressed this gap by comparing the learning outcomes across VR, phone, and mobile learning on two dimensions: recall accuracy and recall speed, as well as learners' attitudes. Additionally, this paper attempted to investigate the differences between 5 and 10 min time duration. By inviting a total of 28 participants to the lab experiments on English vocabulary acquisition, this study yielded no statistically significant difference in vocabulary recall accuracy between VR, computer, and mobile learning. Nonetheless, the significant difference in vocabulary recall speed was generated. Besides, the time spent on vocabulary acquisition was significantly shorter in VR than in computer and mobile learning. Additionally, pairwise comparisons revealed that learners' perceptions and attitudes toward SUS resulted in significantly lower scores in VR than in computer- and mobile-based learning. Further, merely two of the nine items on the VR sickness scale demonstrated a significant raise in visual fatigue as well as significant decline in headache between 5 and 10 min learning duration. Future research should anticipate a larger sample size to ensure the validity of research findings and further construct a 3D modeling for the same language learning settings for non-college, primary, as well as secondary school students.