Bin packing is a classic optimization problem with a wide range of applications from load balancing in networks to supply chain management. In this work we study the online variant of the problem, in which a sequence of items of various sizes must be placed into a minimum number of bins of uniform capacity. The online algorithm is enhanced with a (potentially erroneous) prediction concerning the frequency of item sizes in the sequence. We design and analyze online algorithms with efficient tradeoffs between consistency (i.e., the competitive ratio assuming no prediction error) and robustness (i.e., the competitive ratio under adversarial error), and whose performance degrades gently as a function of the prediction error. This is the first theoretical study of online bin packing in the realistic setting of erroneous predictions, as well as the first experimental study in the setting in which the input is generated according to both static and evolving distributions. Previous work on this problem has only addressed the extreme cases with respect to the prediction error, has relied on overly powerful and error-free prediction oracles, and has focused on experimental evaluation based on static input distributions.