All-in-laterite concrete, in which lateritic material supplies both aggregate fractions, is a low-cost indigenous alternative to conventional concrete, but reinforced use is resisted because its strength variability makes structural performance unpredictable. This paper tests that objection for one laterite source and beam configuration. A published strength regression for this concrete from Idah, Kogi State, Nigeria (twelve mixes, forty-eight standard-cured cubes) was reproduced and embedded in a probabilistic resistance model, with cube-to-in-place strength transfer carried as a random variable. Strength variance used an assumed bounding model of predictive error, batching scatter and ACI 214R-11 control tiers. Fifteen basic variables followed the JCSS Probabilistic Model Code, with Gaussian-copula dependence. The first-order reliability method, verified by importance sampling, assessed sections designed to EN 1992-1-1 and BS 8110 against four limit states. The material supports strength class C12/15. The reference beam attained flexural reliability indices of 5.018 and 5.149 against a target of 3.8; the code-compliant shear-reinforced member attained 8.763, with links governed by minimum-ratio and maximum-spacing rules rather than strength, and 3.297 for quasi-permanent deflection. Concrete strength carried 0.0022 of the flexural limit-state variances, against 0.6332 for imposed load, and the flexural index differed by only 0.029 from conventional concrete of identical class. A concrete-only shear mechanism, screened for an unlinked member, returned 3.103. The flexural conclusion holds while transferred-resistance model uncertainty stays below a coefficient of variation of about 0.15. Strength variability therefore does not govern flexural reliability here, and raising the concrete partial factor would be an ineffective safeguard.