Background <p>Aging reduces physical function, making gait speed a key health indicator. Assessing gait speed using the 4-meter gait speed test (4MGS) requires examining reliability and learning effects across adult and older adult populations.</p> Aims <p>This study determined and compared the within-day test–retest reliability and learning effect of the 4MGS at usual and maximal speeds in adults (18–64&#xa0;years) and older adults (≥ 65&#xa0;years).</p> Methods <p>A total of 287 participants (62.4% female; mean age 47.8 (19.5) years; 216 adults and 71 older adults) were recruited in the community and completed three repetitions of each 4MGS modality (usual and maximal speed). Reliability and measurement error were assessed using intraclass correlation coefficients (ICC_2,1), standard error of measurement (SEM) and minimal detectable change at 95% confidence interval (MDC95). The learning effect was explored using the repeated-measures ANOVA.</p> Results <p>Both age groups demonstrated good-to-excellent reliability across modalities (ICC = 0.90–0.96), with low SEM (4–5%) and MDC95 (11–13%). Small learning effects (&lt; 3%) were identified and significant differences were observed, particularly in the third test, in adults at usual speed and older adults at maximal speed.</p> Conclusions <p>4MGS appears to be a reliable test with an acceptable measurement error in adults and older adults. Although a small learning effect was observed, changes remained within the measurement error, highlighting the importance of interpreting results considering measurement error to reflect actual performance.</p>