Open access academic lectures are potential sources for incidental vocabulary learning. These lectures are available in various formats (transcripts, audios, videos, and video with captions), but no studies have compared the learning of vocabulary in these lectures through different input modes. This study adopted a pretest-posttest design to compare learning at the meaning recall level of 50 words in the same academic lecture through five input modes: reading, listening, reading while listening, viewing, and viewing with captions. One hundred sixty-five English for Academic Purposes learners in China were assigned to five experimental groups and a control group. The experimental groups received the treatment with the assigned input mode while the control group received no treatment. Results show that although learning occurred through all input modes, only viewing significantly contributed to the learning gains. Frequency of occurrence and type of vocabulary significantly predicted the learning gains, but the type of verbal elaboration and nonverbal elaboration did not. This study provides further insights into the value of academic lectures for incidental vocabulary learning and supports the multimedia learning theory and its principles.