The study aimed to investigate the effects of multimedia glosses on Saudi EFL students’ vocabulary learning and long-term retention. It examined which combination of multimedia glosses (textual through L2 definitions, aural input, or visual through pictures) is most effective for Saudi students’ vocabulary learning and long-term retention. It also explored Saudi students’ attitudes and perceptions toward the different multimedia glosses. A quasi-experimental design using pretests, immediate and delayed posttests, and a questionnaire was used to address the questions of the study. One hundred and twenty intermediate-level language learners formed one control and three experimental groups with 30 students in each group. The control group received no glossing instruction. The participants in the three experimental groups received all gloss modes but in different orders. The results of the study showed an overall statistically significant difference between the performance of the experimental and control group students on the immediate and delayed post-tests in favor of the experimental groups’ performance. The results also clarified that both dual glossing modes L2 definition and picture glossing (TV) and L2 definition and audio glossing modes (TA) were significantly better than the single mode L2 definition alone mode (T) in most test sessions. The findings revealed that among the dual glossing modes, the presentation of TV led to better word learning and retention than TA. The findings of the questionnaire showed that the combination of TV was rated as the most effective and desired glossing mode, followed by the combination of TA and text-definition alone mode as the least effective or desired mode. The study concluded with recommendations for students, teachers, material developers, and course designers as well as suggestions for further research.