Cultivated tomato has a narrow germplasm base because of several population bottlenecks in the form of founder events, as well as natural and artificial selections that occurred during domestication and evolution of modern cultivars. The F2 of commercial single-cross hybrids, as well as locally well-adapted varieties, provide germplasm for developing recombinant lines and exploiting genetic variability, respectively. The present study aims to discriminate the breeding value of tomato source material, i.e. commercial hybrids or well-adapted varieties, by (i) estimating tolerance to inbreeding of hybrids or estimating heterosis of diallel hybrids between varieties, (ii) determining undesirable traits and (iii) determining general combining ability (GCA) and specific combining ability (SCA) effects from diallel crosses between hybrids and between varieties. Two hybrids and four varieties were assessed. One hybrid showed 0·03 inbreeding vigour, which was not combined with undesirable traits in the F2 generation. However, negative GCA and positive SCA values did not support the hybrid as source material, provided that hybrids with low inbreeding depression, positive GCA and negative SCA correspond to an F2 capable of developing recombinant lines. The assessment of the varieties showed positive GCA and 0·34 heterosis in one variety, indicating agreement between yield and GCA, and that high-yielding varieties may produce high-yielding hybrids. In conclusion, the proposed mating design, taking into account the tolerance to inbreeding for hybrids and the heritability of general worth for both resources, provides a mechanism for ensuring continued improvement in plant performance through plant selection programmes.