Dynamical systems are useful for bridging discrete and continuous aspects of speech. In this paper, we compare the ability of two models, critically-damped oscillators and General Tau Theory, to predict gestural landmarks.Predictions of the two models were compared with each other and with original kinematic data. The data consisted of electromagnetic articulography recordings of Tibetan collected as part of Geissler (2021). In addition to the landmarks-based approach, this study also uses a a language with a different phonological typology. As compared to results from kinematic thresholds, the critically-damped oscillator model tended to predict that landmarks would take place earlier in time and closer to the target. The General Tau model generally predicted that landmarks would take place later and farther from the target. These results highlight the differences in, and invite further comparison on, the trajectory shape generated by the two models.