The performance of the fully automated BACTEC MGIT 960 (Becton Dickinson, Cockeysville, Md.) system for susceptibility testing of Mycobacterium tuberculosis to critical concentrations of 4 major antituberculosis drugs [streptomycin (STR), isoniazid (INH), rifampin (RIP) and ethambutol (EMB)] was evaluated with 100 clinical isolates and compared to that of the radiometric BACTEC 460 TB (Becton Dickinson, Sparks, Md.) system. The results of BACTEC 960 MGIT, which are not compatible with BACTEC 460 TB, were re-evaluated according to the results obtained. The discordance was 2.25% at the critical concentrations of both systems. After the reproducibility testing, the discordance between 2 systems reduced to 1.25% and the concordance increased to 98.75%. In our study, no very major error was observed; major error was detected as 2.25% (STR 2, INH 4, RIF 2 and EMB 1) in the 1(st) trial, and 1.25% (INH 3, RIF 2) in the 2(nd) trial. When BACTEC 460 TB has been compared with BACTEC 960 MGIT, the overall specificity was 100% in all tests, the sensitivity was 95.6-98.9% in the initial testing, however, the sensitivity changed to 96.7-100% after the reproducibility testing. It was concluded that there was no statistically significant disconcordance between the two test systems. Thus BACTEC MGIT 960 might be an alternative to BACTEC 460 TB for susceptibility testing of antituberculosis agents.