Live trapping is an invaluable and commonly used technique for studying small mammal populations. There are several types of commercially available live-traps, but some models have been shown to differ in terms of species-specific efficiency. Such differences could lead to biased results, and knowledge of such bias is important when comparing results of different studies. The Ugglan Special live-trap is one commonly used type, and it is available in several different models. I studied performance of the two most common models (No. 1 and No. 2) to find out whether they differ in terms of overall efficiency and body-mass-dependent efficiency, when trapping small rodents. I used 48 traps of each model in boreo-nemoral forest during three years, which amounted to 3456 trap nights, during which 268 bank voles (Myodes glareolus) (26%) and wood mice (Apodemus sylvaticus) (74%) were captured. There was no difference in body mass between the animals captured by the two trap models. Model No. 2, however, was more efficient in capturing bank voles, but not wood mice, than model No. 1. This difference may have been caused by the difference in the entrance design. This species-specific dissimilarity in trap efficiency should be taken into account when using Ugglan Special live-traps in ecological research relying on mark–recapture methods. Additional studies are needed to see if other species of small mammals are affected in similar ways and if the differences are consistent among habitat types.