In vivo relative bioavailability (RBA) assessments were conducted for cobalt (Co) and nickel (Ni), in baghouse dust, surface dust, and soil collected in a neighborhood near a metal forge. Baghouse dust contained 8,000 mg/kg Co and 45,000 mg/kg Ni; surface dust contained 970 mg/kg Co and 6000 mg/kg Ni; and soil contained 53 mg/kg Co and 330 mg/kg Ni. Test materials (baghouse dust, surface dust, and soil) and reference materials-cobalt chloride (CoCl2) and nickel sulfate (NiSO4) -were administered daily for 14 or 21 days, twice daily, to juvenile swine. Daily intake doses for Co were ≤229 mg/kg-day and for Ni ≤1,419 mg/kg-day. Concentrations of Co and Ni were measured in liver, kidneys, femur, and duodenum; 24-hour urinary excretion of each metal was also measured. Multiple linear regression modeling was performed for tissue concentration or urinary excretion vs. dose in each material, with weighting as inverse variance in each dose group. RBA for each test material was estimated as the ratio of best-fit slopes for test material and reference material. Models were evaluated on residual heteroscedasticity and goodness-of-fit metrics. Uncertainty in RBA was estimated from uncertainty in the best-fit slopes. Cobalt RBA in baghouse dust was 8.7% (90% CI: 2.6%-15%); in surface dust, 12% (90% CI: 0.47%-23%); and in soil, 3.8% (90% CI: -6.0%-14%). Nickel RBA in baghouse dust was 4.5% (-57% to 53%); in surface dust, 2.4% (90% CI: – 9.6% to 14%); and in soil, 20% (90% CI: -78% to 106%). Nickel RBA estimates were highly uncertain in soil and baghouse dust, likely because the soluble reference material (NiSO4) was poorly absorbed. These data demonstrate that cobalt and nickel in soil and dust affected by metal alloys are resistant to absorption. This study provides useful information to be incorporated into site-specific human health risk assessments.