Aim Key to the successful management of paediatric pulmonary tuberculosis (PTB) lies in the early detection and proper treatment. We evaluated the performances of modern diagnostic tests: loop‐mediated isothermal amplification (LAMP‐IS6110), Xpert MTB/RIF (Cepheid) and mycobacteria growth indicator tube (BACTEC MGIT 960 culture) against a modified version of international consensus diagnostic definition (i.e. composite reference standard (CRS)). Methods A cross‐sectional analytical study was conducted in a tertiary care hospital in North India from July 2016 to December 2017 involving 100 children <14 years with suspected PTB. Respiratory specimens (sputum, gastric lavage and/or bronchoalveolar lavage) were collected and subjected to LAMP‐IS6110, Xpert MTB/RIF and BACTEC MGIT 960 culture assay. Results Fifty‐five children had confirmed and probable TB according to the CRS (prevalence = 58.5%). The sensitivity of BACTEC MGIT 960 culture, Xpert MTB/RIF and LAMP‐IS6110 assay was 14%, 9.1% and 10.91%, respectively, when compared against the predefined CRS. The specificity for all these tests was 100%. When compared with BACTEC MGIT 960 culture as the gold standard, the LAMP‐IS6110 assay and Xpert MTB/RIF assay had the sensitivity of 85.71% (95% CI: 42.13–99.64%) and 71.43% (95% CI: 29.04–96.33%), respectively. The specificity of both assays was 100%. Conclusions We noted that LAMP‐IS6110 performed better than Xpert MTB/RIF (Cepheid) in terms of sensitivity when compared against BACTEC MGIT 960 culture as reference standard, though specificity of both the tests was comparable. The diagnostic performance of BACTEC MGIT 960 culture was better than LAMP‐IS6110 and Xpert MTB/RIF in paediatric PTB, when compared against CRS.