For over a century, various neuroanatomical measures have been employed as assays of cognitive ability in comparative studies. Nevertheless, it is still unclear whether these measures actually correspond to cognitive ability. A recent meta-analysis of cognitive performance of a broad set of primate species has made it possible to provide a quantitative estimate of general cognitive ability across primates. We find that this estimate is not strongly correlated with neuroanatomical measures that statistically control for a possible effect of body size, such as encephalization quotient or brain size residuals. Instead, absolute brain size measures were the best predictors of primate cognitive ability. Moreover, there was no indication that neocortex-based measures were superior to measures based on the whole brain. The results of previous comparative studies on the evolution of intelligence must be reviewed with this conclusion in mind.