In light of South Africa’s high prenatal HIV prevalence and infant mortality rate, a cluster randomized controlled trial was conducted to evaluate an intervention called Philani+, which used community health workers (known as Mentor Mothers) to deliver pre- and postnatal home visits in Cape Town, South Africa, to improve maternal and child health. We assessed the costs and benefits of this intervention and made comparisons with other scenarios that depicted increased capacity and provision of nurse-delivered care. The recurrent cost of the twenty-four-month intervention was US$80,001. The major health outcomes analyzed were differences in the proportion of infants who were low birthweight, stunted, and suboptimally breastfed between intervention and control groups. Each case of low birthweight averted cost US$2,397; of stunted growth, US$2,454; and of suboptimal breastfeeding, US$1,618. Employment of community health workers was cost saving compared to that of nurses. Philani+ improved child health at a relatively low cost, considering the health system costs associated with low birthweight and undernutrition. The model could be suitable for replication in low-resource settings to improve child health in other countries.