Children identified as cognitively gifted, in comparison with age-matched mainstream samples, are advantaged in numerous areas, including mathematics, speed and efficiency in cognitive processing, and resistance to interfering stimuli. Although working memory (WM) has been implicated as a factor mediating these advantages, evidence suggests that gifted children may not be advantaged in all aspects of WM function. We hypothesized that this difference is related to the contrast between mental (related to prefrontal dopamine circuits) and perceptual attention (likely related to prefrontal acetylcholine circuits). Specifically, it was expected that cognitively gifted children would excel in WM tasks taxing mental but not perceptual attention. Ninety-one children from grades 4 and 8, in the gifted and mainstream academic streams, received WM tasks requiring primarily perceptual attention (SOPT) and mental attention (n-back), as well as measures of mental-attentional capacity, shifting, and inhibition. Gifted children outperformed their mainstream peers on all tasks, except SOPT (even when mental demand was matched). Results demonstrate a necessary distinction between mental and perceptual attention in the measurement of WM.