The Keystroke-Level Model (KLM) is a model for predicting the execution time of routine tasks. Initially, it had been devised for standard keyboard-desktop settings but an extension of this model for interactions with mobile phones has been described by Holleis et al. [10]. We propose a considerable update of this KLM focusing on NFC-based applications and interactions which are continuously gaining interest. Insufficiencies within the previous model regarding operators for Pointing, Mental Acts, and System Response Time are treated. We present the results of several studies conducted in order to update the values of these operators. A specific focus is put on the differences between static (NFC tags behind a printed poster or object) and dynamic interfaces (tagged displays or projections). Finally, we validate our results by modeling two applications with the former and the proposed model. The latter performed consistently better when compared with measurements from real user interaction data.