Our studies of children in a rural Bangladeshi area, with varying concentrations of arsenic in well-water, indicated modest impact on child verbal cognitive function at 5 years of age.Follow-up of arsenic exposure and children’s cognitive abilities at school-age.In a nested sub-cohort of the MINIMat supplementation trial, we assessed cognitive abilities at 10 years of age (n = 1523), using Wechsler Intelligence Scale for Children (WISC-IV). Arsenic in maternal urine and erythrocytes in early pregnancy, in child urine at 5 and 10 years, and in hair at 10 years, was measured using Inductively Coupled Plasma Mass Spectrometry.Median urinary arsenic at 10 years was 58 µg/L (range 7.3–940 µg/L). Multivariable-adjusted regression analysis showed that, compared to the first urinary arsenic quintile at 10 years (<30 µg/L), the third and fourth quintiles (30–45 and 46–73 µg/L, respectively) had 6–7 points lower Full developmental raw scores (B: −7.23, 95% CI −11.3; −3.18, and B: −6.37, 95% CI −10.5; −2.22, respectively), corresponding to ~0.2 SD. Verbal comprehension and Perceptual reasoning seemed to be affected. Models with children’s hair arsenic concentrations showed similar results. Maternal urinary arsenic in early pregnancy, but not late pregnancy, showed inverse associations with Full developmental scores (quintiles 2–4: B: −4.52, 95% CI −8.61; −0.43, B: −5.91, 95% CI −10.0; −1.77, and B: −5.98, 95%CI −10.2; −1.77, respectively, compared to first quintile), as well as with Verbal comprehension, Perceptual reasoning, and Processing speed, especially in girls (p < 0.05 for interaction of sex with Full developmental scores and Perceptual reasoning). In models with all exposure time points included, both concurrent exposure at 10 years and early prenatal exposure remained associated with cognitive abilities.Both early prenatal and childhood arsenic exposure, even at low levels (about 50 µg/L in urine), was inversely associated with cognitive abilities at school-age, although the estimates were modest.