A pan-tuberculosis regimen that could be initiated without knowledge of drug susceptibility has been proposed as an objective of tuberculosis regimen development. We modelled the health and economic benefits of such a regimen and analysed which of its features contribute most to impact and savings. We constructed a mathematical model of tuberculosis treatment parameterised with data from the published literature specific to three countries with a high tuberculosis burden (India, the Philippines, and South Africa). Our model simulated cohorts of newly diagnosed tuberculosis patients, including drug susceptibility testing if performed, regimen assignment, discontinuation, adherence, costs, and resulting outcomes of durable cure (microbiological cure without relapse), need for retreatment, or death. We compared a pan-tuberculosis regimen meeting the WHO 2023 target regimen profile against the standard of care of separate rifampicin-susceptible and rifampicin-resistant regimens. We estimated incremental cures; averted deaths, secondary cases, and costs; and prices below which a pan-tuberculosis regimen would be cost saving. We also assessed scenarios intended to describe which mechanisms of benefit from a pan-tuberculosis regimen (including improved characteristics compared with the current rifampicin-susceptible and rifampicin-resistant regimens and improved regimen assignment and retention in care for patients with rifampicin-resistant tuberculosis) would be most impactful. Results are presented as a range of means across countries with the most extreme 95% uncertainty intervals (UIs) from the three UI ranges. Compared with the standard of care, a pan-tuberculosis regimen could increase the proportion of patients durably cured after an initial treatment attempt from 69-71% (95% UI 57-80) to 75-76% (68-83), preventing 30-32% of the deaths (20-43) and 17-20% of the transmission (9-29) that occur after initial tuberculosis diagnosis. Considering savings to the health system and patients during and after the initial treatment attempt, the regimen could reduce non-drug costs by 32-42% (22-49) and would be cost saving at prices below US$170-340 (130-510). A rifamycin-containing regimen that otherwise met pan-tuberculosis targets yielded only slightly less impact, indicating that most of the benefits from a pan-tuberculosis regimen resulted from its improvements upon the rifampicin-susceptible standard of care. Eliminating non-adherence and treatment discontinuation, for example via a long-acting injectable regimen, increased health impact and savings. In countries with a high tuberculosis burden, a shorter, highly efficacious, safe, and tolerable regimen to treat all tuberculosis could yield substantial health improvements and savings. Bill & Melinda Gates Foundation.