Tuberculosis remains a major source of morbidity and mortality worldwide, with 10 million cases and 1.5 million deaths in 2018. Achieving 'End TB' prevention and care goals by 2035 will likely require a new tuberculosis vaccine. The tuberculosis vaccine development pipeline has seen encouraging progress; however, questions around their population impact and implementation remain. Mathematical modelling investigates these questions to inform vaccine development and deployment strategies. We provide an update on the current vaccine development pipeline, and a systematic literature review of mathematical modelling of the epidemiological impact of new tuberculosis vaccines. Fourteen prophylactic tuberculosis vaccine candidates are currently in clinical trials. Two candidates have shown promise in phase II proof-of-concept efficacy trials: M72/AS01E demonstrated 49.7% (95% CI; 2.1, 74.2) protection against tuberculosis disease, and BCG revaccination demonstrated 45.4% (95% CI; 6.4, 68.1) protection against sustained Mycobacterium tuberculosis infection. Since the last modelling review, new studies have investigated the epidemiological impact of differential vaccine characteristics, age targeting and spatial/risk group targeting. Critical research priorities for M72/AS01E include completing the currently in-design trial, powered to improve the precision of efficacy estimates, include uninfected populations and further assess safety and immunogenicity in HIV-infected people. For BCG revaccination, the priority is completing the ongoing confirmation of efficacy trial. Critical modelling gaps remain on the full value proposition of vaccines, comparisons with other interventions and more realistic implementation strategies. Using carefully designed trials and modelling, we must prepare for success, to ensure that new vaccines will be promptly received by those most in need.