The use of genetic distances to identify species within the framework of DNA barcoding has to some extent improved the development of biodiversity studies. However, using a fixed empirical threshold to delimit species may lead to overestimating species diversity. In this study, we use a new data set of COI sequences for 366 specimens within the genus of Cletus as well as conduct an analysis on the same genetic data for collected morphologically defined species from previous phylogeographical studies, to test whether high intraspecific genetic divergences are common with the premises of comprehensive sampling. The results indicate C. graminis Hsiao & Cheng 1964, is the same species with C. punctiger (Dallas, 1852) and should be synonymized and that the distributional record of C. pugnator (Fabricius, 1787) in China is correct. High intraspecific genetic differentiations (0%–4.35%) were found in C. punctiger. Furthermore, as to the mined data, the maximum intraspecific K2P distances of 186 species (48.44% of 384) exceed 3%, and 101 species (26.30%) can be divided into two or more clusters with a threshold of 3% in cluster analysis. If genetic distance is used to delimit species boundaries, the minimum interspecific K2P distance of the congeneric species should be considered rather than only using the fixed empirical value; otherwise, the species richness may be overestimated in some cases.