Introduction: The aim of this international project was to establish a species-specific Clinical Breakpoint for interpretation of Antimicrobial Susceptibility Testing of benzylpenicillin (BP) in horses. Methods: A population pharmacokinetic model of BP disposition was developed to compute PK/PD cutoff values of BP for different formulations that are commonly used in equine medicine around the world (France, Sweden, USA and Japan). Investigated substances were potassium BP, sodium BP, procaine BP, a combination of procaine BP and benzathine BP and penethamate, a prodrug of BP. Data were collected from 40 horses that provided 63 rich profiles of BP corresponding to a total of 1022 individual BP plasma concentrations. Results: A 3-compartment disposition model was selected. For each of these formulations, the PK/PD cutoff was estimated for different dosage regimens using Monte Carlo simulations. The fAUC/MIC or fT>MIC were calculated with a free BP fraction set at 0.4. For fAUC/MIC, a target value of 72 h (for a 72h treatment) was considered. For fT>MIC, efficacy was assumed when free plasma concentrations were above the explored MIC (0.0625-2 mg/L) for 30 or 40 % of the dosing interval. For continuous infusion, a fT>MIC of 90 % was considered. It was shown that a PK/PD cutoff of 0.25 mg/L can be achieved in 90 % of horses with routine regimen (typically 22,000 IU/kg or 12.4 mg/kg per day) with IM procaine BP once a day (France, Japan, Sweden but not USA1) and with IM sodium BP at 14.07 mg/kg, twice a day or IV sodium BP infusion of 12.4 mg/kg per day. In contrast, penethamate and the combination of procaine BP and benzathine BP were unable to achieve this PK/PD cutoff not even an MIC of 0.125 mg/L. Discussion: The PK/PD cutoff of 0.25 mg/L is one dilution lower than the clinical breakpoint released by the CLSI (0.5 mg/ L). From our simulations, the CLSI clinical breakpoint can be achieved with IM procaine BP twice a day at 22,000 IU i.e. 12.4 mg/kg.