Responding to varying customer needs for product customization, quantities, or lead times requires increased manufacturing flexibility. In certain contexts, the reverse is true, i.e., customer requirements are somewhat flexible, providing additional options in production planning. This paper considers such an opportunity faced by a specialty steel manufacturer whose customers allow flexibility in product specifications. We develop and test a model and solution methodology to exploit this flexibility during production planning. The model also applies to other situations in which manufacturers can select, within limits, the size or quantities of items produced using limited capacity. We formulate the problem as a profit-maximizing mixed-integer program with an interesting embedded packing subproblem using expandable items. We develop a composite solution method that combines model enhancement using strong valid inequalities, Lagrangian relaxation, and heuristic approaches. Computational results using real data from a specialty steel manufacturer and randomly generated test problems show that the algorithm generates solutions that are, on average, within 0.59% of optimality. For the problem instances based on real data, the model increases contribution to profit by over 7% relative to current practice at the steel manufacturing facility.