Traditional vegetables are very important for food and nutrition security in sub-Saharan Africa but have not received much attention in crop improvement research. One exception is amaranth (Amaranthus spp.) for which international breeding research began at the World Vegetable Center in Tanzania in 2004. This study is the first impact evaluation of this research program. It uses a representative sample of 1,355 amaranth-producing households from five agroecological zones of Tanzania to collect questionnaire-based data and plant material. The genetic identity of the amaranth cultivars used by the farmers was identified using DNA fingerprinting, while the impact on crop yield, production cost and sales was quantified using propensity score matching. The analysis shows that 66% of Tanzania’s cultivated area under amaranth uses cultivars derived from the amaranth breeding program. These cultivars had a mean vegetable yield that was 6.1 ton/ha higher (+ 48%; p = 0.002) than that of other cultivars, while there was no significant impact on the cost of production or quantity sold. This suggests tremendous impact of genetic improvement research in traditional vegetables.