In this paper, we explore the trade-offs between sensors' and models' accuracy for state estimation in battery management systems. If, in a battery pack, high quality sensors were used, then state estimation (or monitoring) would be improved at the expenses of hardware costs. On the other hand, if accurate models were used within the estimation algorithms, better estimates could be produced at the expenses of engineering time required for modeling the battery and parameterizing the model, as well as the inevitable increase in microprocessors' power due to the increase in the algorithm's computational complexity. Hence, the research question we ask is: for a given budget or a given estimation error tolerance, what is the minimum sensor accuracy and model accuracy needed in order to achieve that target? In its simple form, this is a two-dimensional multi-objective optimization problem.