This study explores the impact of instructional modality and session length on student success and retention in first-semester Japanese and Spanish language courses at a large urban community college. The research analyzed pre- and post-tests, final grades, and retention rates across multiple academic terms, including standard 16-week and compressed 8-week sessions during the fall and spring semesters, as well as intensive 10-week and 5-week terms offered in the summer. Courses were delivered in asynchronous, synchronous, hybrid, hyflex, and face-to-face formats. Instructor perspectives, gathered through interviews, recorded conversations, and email correspondence, provided anecdotal insights into instructional practices and student engagement. Drawing on Vygotsky's Zone of Proximal Development, Krashen's Input Hypothesis, and Long's Interaction Hypothesis, the findings reveal that Japanese I students performed best in 16-week sessions--particularly in asynchronous formats--where extended input and longer exposure supported gains in post-test scores and retention. Spanish I students showed stronger performance during the first half of compressed terms, especially in hybrid and synchronous formats. In contrast, outcomes were consistently lower in the second-session short terms. Across both languages, courses featuring real-time instructor interaction resulted in higher success rates than asynchronous formats. These findings suggest that effective language instruction depends not only on session duration, but also on how well instructional design aligns with learners' developmental needs, the structure of the academic calendar, and the inherent complexity of the target language.