Space-based nulling interferometry is one of the most promising solutions to spectrally characterize the atmosphere of rocky exoplanets in the mid-infrared (3 to 20 μm). It provides both high angular resolution and starlight mitigation. This observing capability depends on several technologies. A CubeSat (up to 20 kg) or a medium satellite (up to a few hundreds of kg), using a Bracewell architecture on a single spacecraft could be an adequate technological precursor to a larger, flagship mission. Beyond technical challenges, the scientific return of such a small-scale mission needs to be assessed. We explore the exoplanet science cases for various missions (several satellite configurations and sizes). Based on physical parameters (diameter and wavelength) and thanks to a state-of-the-art planet population synthesis tool, the performance and the possible exoplanet detection yield of these configurations are presented. Without considering platform stability constraints, a CubeSat (baseline of b ≃ 1 m and pupils diameter of D ≃ 0.1 m) could detect ≃7 Jovian exoplanets, a small satellite (b ≃ 5 m∕D ≃ 0.25 m) ≃120 exoplanets, whereas a medium satellite (b ≃ 12.5 m∕D ≃ 0.5 m) could detect ∼250 exoplanets including 51 rocky planets within 20 pc. To complete our study, an analysis of the platform stability constraints (tip/tilt and optical path difference) is performed. Exoplanet studies impose very stringent requirements on both tip/tilt and OPD control.