For motion adaptive radiotherapy, dynamic multileaf collimator (dMLC) tracking can be employed to reduce treatment margins by steering the beam according to the organ motion. Until now, the Elekta Agility 160 MLC has hitherto not been evaluated for its tracking suitability under physiologic conditions. Both dosimetric performance and latency are key figures and need to be assessed generically, independent of the used motion sensor. In this paper, we propose to use harmonic functions directly fed to the MLC to determine its latency during continuous motion. Furthermore, a control variable is extracted from a camera system and fed to the MLC. Using this setup, film dosimetry and subsequent γ statistics are performed, evaluating the response when tracking magnetic resonance imaging (MRI)-based physiologic motion in a closed-loop. The delay attributed to the MLC itself was shown to be a minor contributor to the overall feedback chain as compared to the impact of imaging components such as MRI sequences. Delay showed a linear phase behaviour of the MLC employed in continuously dynamic applications, which enables a general MLC-characterization. Using the exemplary feedback chain, dosimetry showed a vast increase in pass rate employing γ statistics. In this early stage, the tracking performance of the Agility using the test bench yielded promising results, making the technique eligible for the translation to tracking using clinical imaging modalities.