Rapid and accurate diagnosis of symptomatic patients of pulmonary tuberculosis (TB) is highly desirable to minimize the spread of the disease in the society. We, therefore, compared the usefulness of various conventional diagnostic methods, the in-house polymerase chain reaction (PCR), and the FASTPlaque assay in this study. Laboratory data of 150 patients with clinical diagnosis of pulmonary TB and 50 controls were included in this study. The sputa from all these 200 individuals were subjected to acid-fast staining, culture on Lowenstein-Jensen (L-J) slants, automated BACTEC-MGIT-960 culture methods, and a mycobacteriophage assay. A mycobacterium genus and Mycobacterium tuberculosis species-specific PCRs were also done and samples positive on both PCRs were considered as standard for comparison. Of the 5 in vitro diagnostic tests, PCR method was found to be the most rapid, sensitive, and specific, detecting all the 150 cases of pulmonary TB without any false-positive and negative result. In comparison with PCR the sensitivity of MGIT-960 was 90%, followed by FASTPlaque assay (76.7%), L-J culture method (73.3%), and microscopy (60%). The mean detection time for smear-positive and smear-negative samples was 12.5 and 14 days in MGIT-960 and 18 and 25 days for L-J method, respectively. The FASTPlaque failed to detect mycobacteria from the paucibacillary samples. The contamination rates for MGIT-960, L-J, and FASTPlaque assays were 4, 8 and 10%, respectively. The best correlation with mycobacterial load in the specimen was observed in BACTEC-MGIT-960 showing 66.6% detection rate in paucibacillary, 83.3% in 1+ samples, and 100% in 2+ and 3+ samples. Out of the 150 patients, 140 (93.3%) could be diagnosed by one or more nonmolecular methods. Therefore, it could be concluded that combination of three or more in vitro diagnostic methods will have acceptable detection level.