Introduction: The worrying increase in multi-resistant Gram-negative bacilli (GNB) infections and the glaring lack of therapeutic alternatives has allowed colistin to regain a place in the antibiotic arsenal, often as a last resort. This study aimed to evaluate a method for testing sensitivity to colistin that is simple, inexpensive, and accessible: disk diffusion, compared to the gold standard method of liquid microdilution. Materials and methods: A comparative study was conducted on 124 strains of Gram-negative bacilli. Each strain was tested by agar diffusion (50 µg colistin disk) and broth microdilution (Sensititre® plate). Results: Our results showed that the diffusion method does not meet the performance criteria established by the CLSI, with a particularly high major error rate (45.8%) and a categorical agreement of 83.1%. Conclusion: Disk diffusion cannot be considered a reliable method for testing colistin susceptibility in the laboratory.