Colistin is the last resort treatment against resistant Gram-negative bacteria, necessitating reliable and rapid means for sensitivity testing of colistin. Automated-systems like VITEK®2 is adopted to determine minimum inhibitory concentration (MIC) due to its easy usage. Broth microdilution (BMD) for colistin MIC was suggested by EUCAST and CLSI. The objective was to compare and evaluate colistin MIC by BMD and VITEK®2 against Gram-negative organisms from ICU in a tertiary care hospital. Clinically significant organisms isolated from ICU patients were included. MIC was determined using BMD and VITEK®2. Very major error (VME), major error (ME), essential agreement (EA), categorical agreement (CA), positive-predictive value (PPV), negative-predictive value (NPV), sensitivity and specificity were analysed. 533 isolates were obtained from blood (435,81.60%), respiratory samples (57,10.70%), pus and exudates (20,3.80%), urine (18,3.40%) and CSF (3,0.60%). The Enterobacterales were K. pneumoniae (185,34.70%), E. coli (73,13.70%) and E. cloacae (26,4.90%) while non-fermenters were A. baumannii (209,39.20%) and P. aeruginosa (40,7.50%). The VITEK®2 sensitivity was >99%; specificity ranged from 14.28-52.94%. PPV was 93.81% while NPV was 93.75%. VME ranged from 47-100% between isolates. ME was up to 20%. Highest VME was obtained in E. coli (100%). The total EA and CA observed was 68.5% and 99.79% respectively. VITEK®2 failed to detect the resistance in 32(60%) isolates. Obtained VME and ME values were >3%, which is unacceptable as per the standard guidelines. EA of ≥90% wasn’t obtained. Sensitivity for VITEK®2 was >99%, but had low specificity (14.28%). Hence, VITEK®2 is not reliable for colistin susceptibility testing.