Whether PD‐L1‐positive patients derive more overall survival benefit from PD‐1/PD‐L1 inhibitors in the treatment of advanced solid tumours is unclear. We systematically searched the PubMed, Cochrane library and EMBASE databases from January 1, 1966 to March 1, 2019, to identify randomised controlled trials of PD‐1/PD‐L1 inhibitors (nivolumab, pembrolizumab, atezolizumab, durvalumab and avelumab) that had available hazard ratios (HRs) for death according to PD‐L1 status. A random‐effects model was used to calculate the pooled overall survival (OS) HR and 95% CI among PD‐L1‐positive and PD‐L1‐negative patients. An interaction test was performed to evaluate the heterogeneity between the two estimates. A total of 24 randomised trials, involving 12,966 participants, fulfilled the inclusion criteria. An OS benefit of PD‐1/PD‐L1 inhibitors was found in both PD‐L1‐positive patients (HR, 0.65; 95% CI, 0.60–0.70) and PD‐L1‐negative patients (HR, 0.82; 95% CI, 0.74–0.91) even at the minimum cut‐off value of 1%. Significant differences in the efficacy of PD‐1/PD‐L1 inhibitors between PD‐L1‐positive and PD‐L1‐negative patients were noted at different cut‐off values. Moreover, there was a positive dose–response relationship between PD‐L1 positivity and OS benefit (HR for 1%, 0.58, [0.50, 0.67]; 5%, 0.52 [0.43, 0.64]; 10%, 0.50 [0.40, 0.63]). Subgroup analyses showed that these results were generally consistent, regardless of study design, line of treatment, treatment type, tumour type, PD‐L1 staining cell type and median follow‐up time. We demonstrated that PD‐1/PD‐L1 inhibitors significantly improved OS in both PD‐L1 positive and PD‐L1 negative patients compared to controls, but the magnitude of benefit was clinically PD‐L1‐dependent.