It is suggesting a hard cutoff since it says "most" attractive, a superlative, instead of "more" attractive. And it is not pareto optimal, because if the models get really good and are scoring 10x more for 10x cheaper that green region will no longer be pareto optimal.