Seeing large or small objects in a choice-reaction task facilitates power and precision grip, respectively. According to the motor-simulation account, this congruity effect reflects an automatic access to object representation, including the grip usually associated to the object. Alternatively, this effect might come from an overlap between magnitude codes of the manipulable objects and those of the response's outcomes. In Experiment 1, participants saw objects usually grasped with power or precision grip and had to mimic these gestures on a specific device. Large-switch responses (e.g., power grip) always produced a high-pitch tone (i.e., a small magnitude perceptual outcome) while small-switch responses (e.g., precision grip) always produced a low-pitch tone (i.e., a large magnitude perceptual outcome). We also manipulated the instructions asking one group to perform specific grips, while the other had to produce specific pitch tones. Even if results revealed shorter RTs in the object-response compatible condition compared to the non-compatible condition for both conditions, they also revealed a decrease of the congruity effect in the tone-instruction group (+ 13 ms) compared to the switch-instruction group (+ 25 ms). We conducted a second experiment based on the tone-instruction group but the response device was installed horizontally. Thus, the grasping component of the responses was removed as in Heurley et al. (2020). Results revealed a congruity effect supporting the magnitude coding account instead of the motor-simulation account. Results also seem to support a kind of permeability of the coding to the influence of intentional processes related to the processing of action outcomes.