Lentil (Lens culinaris Medik.) is an important and expanding crop in southern Australia and a significant crop in western Canada. Currently, production in both countries is limited by an inability to effectively control weeds, due in part to a lack of registered safe and effective herbicides. Metribuzin is a broad-spectrum herbicide providing an alternative weed control option to the imidazolinones, but it has low crop safety in lentil. Two methods, germplasm screening using a hydroponic sand assay and field screening of a large mutated population of the Australian cultivar ‘PBA Flash’ were initially used to identify lines with putative metribuzin tolerance over current cultivars. Dose–response experiments showed the germplasm line SP1333 had GR₅₀ (the rate required to reduce dry weight 50%) values up to four times higher than PBA Flash. However, the mutation selections M043 and M009 had GR₅₀ values more than 25 times higher than PBA Flash. A field study in Canada, under conditions of induced shade and no shade 72 h before POST application of metribuzin, confirmed the intermediate level of tolerance in SP1333 and the high level in the two mutant lines compared with 20 Canadian and Australian genotypes. This relative increase in metribuzin tolerance of the two mutant lines over the parent cultivar is higher than all previous reports in a range of crop species. The development of large mutant populations combined with large M² field screens was a successful method for developing high levels of metribuzin tolerance in lentil. The estimated mutation rate of the mutant lines was 9.4 × 10 ^{− 8}. All three lines are currently being used as parents in lentil breeding programs.