Within crystallization theory, (Matveev's) complexity of a 3-manifold can be estimated by means of the combinatorial notion of GM-complexity. In this paper, we prove that the GM-complexity of any lens space L(p, q), with p ≥ 3, is bounded by S(p, q)−3, where S(p, q) denotes the sum of all partial quotients in the expansion of q p as a regular continued fraction. The above upper bound had been already established with regard to complexity; its sharpness was conjectured by Matveev himself and has been recently proved for some infinite families of lens spaces by Jaco, Rubinstein and Tillmann. As a consequence, infinite classes of 3-manifolds turn out to exist, where complexity and GM-complexity coincide.Moreover, we present and briefly analyze results arising from crystallization catalogues up to order 32, which prompt us to conjecture, for any lens space L(p, q) with p ≥ 3, the following relation: k(L(p, q)) = 5 + 2c(L(p, q)), where c(M ) denotes the complexity of a 3-manifold M and k(M ) + 1 is half the minimum order of a crystallization of M .