ABSTRACT Animal bioassays have been developed to estimate oral relative bioavailability (RBA) of metals in soil, dust, or food for accurate health risk assessment. However, the comparability in RBA estimates from different labs remains largely unclear. Using 12 soil and soil-like standard reference materials (SRMs), this study investigated variability in lead (Pb) and arsenic (As) RBA estimates employing a mouse bioassay in 3 labs at Nanjing University, University of Jinan, and Shandong Normal University. Two performances of the bioassay at Nanjing University in 2019 and 2020 showed reproducible Pb and As RBA estimates, but increasing the number of mouse replicates in 2020 produced more precise RBA measurements. Although there were inter-lab variations in diet consumption rate and metal accumulation in mouse liver and kidneys following SRM ingestion due to differences in diet composition, bioassays at 3 labs in 2019 yielded overall similar Pb and As RBA estimates for the 12 SRMs with strong linear correlations between each 2 of the 3 labs for Pb (R2 = 0.95–0.98 and slope = 0.85–1.02) and As RBA outcomes (R2 = 0.46–0.86 and slope = 0.56–0.79). The consistency in RBA estimates was attributed to the relative nature of the final bioavailability outcome, which might overcome the inter-lab variation in diet consumption and metal uptake in mice. These results increased the confidence of use of mouse bioassays in bioavailability studies.