Purpose: Microbial leakage and colonization between implants and their abutments may cause inflammatory reactions in the peri-implant tissues. This study evaluated microbial leakage at the implant-abutment interface with a new in vitro model. Materials and Methods: Bacterial leakage was tested during dynamic loading in a 2-axis chewing simulator. The authors theorized that dynamic loading would decrease the stability of the implant-abutment connections and thereby lead to bacterial penetration along the gap. Five different implant systems with 8 standard implant-abutment combinations for single molar crowns were tested. The internal aspects of the implants were inoculated with a bacterial suspension and connected to the superstructure with the recommended torque. The specimens were immersed in a nutrient solution and loaded with 1,200,000 cycles of 120 N in the chewing simulator. Results: Statistically significant differences (P = .05) between implant systems with respect to number of chewing cycles until bacterial penetration were found. Discussion: The degree of penetration in a specific implant system presumably is a multifactorial condition dependent on the precision of fit between the implant and the abutment, the degree of micromovement between the components, and the torque forces used to connect them. Conclusion: It was concluded that the newly developed test model is a sensitive tool for the detection of differences between current implant systems with respect to their ability to prevent bacterial penetration at the implant-abutment interface under dynamic loading conditions. Palabras clave: abutments, bacterial leakage, chewing simulation, dental implants