Objective: Azithromycin is an alternative to treat invasive non-typhoidal Salmonella (iNTS) infections. We determined its epidemiological cut-off (ECOFF) and compared azithromycin susceptibility testing methods for iNTS. Methods: We used EUCAST ECOFFinder to determine the minimum inhibitory concentrations (MIC; obtained by broth microdilution) ECOFF and corresponding disk zone diameters of 515 iNTS from blood cultures in Democratic Republic of Congo, Burkina Faso, Rwanda, and Cambodia. Transferable resistance mechanisms were determined by polymerase chain reaction. We compared azithromycin susceptibility testing by semi-automated broth microdilution (customized Sensititre panel; reference), agar dilution, gradient tests (bioMérieux, Liofilchem, HiMedia; read at 80% (MIC80%) and 100% inhibition (MIC100%)), and disk diffusion (Rosco, Oxoid, BD, Liofilchem) for 161 wild- and 198 non–wild-type iNTS. Results: Azithromycin MIC ECOFF was 16 mg/L corresponding to a 12 mm zone diameter; mphA was detected in 192/197 non–wild- and 0/47 wild-type iNTS. Categorical agreement was excellent (≥98%) for all methods. Essential agreement was very good for agar dilution (>90%) but moderate for gradient tests (MIC80%: 52% to 71% and MIC100%: 72% to 91%). Repeatability was good for all methods/brands. Interreader agreement was high for broth microdilution and agar dilution (all ≤1 twofold dilution difference) and disk diffusion (>96% ≤3 mm difference) but lower for gradient tests (MIC80% & MIC100%: 83% to 94% ≤1 twofold dilution difference). Discussion: Azithromycin ECOFF of iNTS was 16 mg/L, i.e. equal to Salmonella Typhi. Disk diffusion is an accurate, precise, and user-friendly alternative for agar dilution and broth microdilution. Reading gradient tests at 100% instead of 80% inhibition improved accuracy and precision.