@Jessica Chan good question — the 50km figure is the drone-to-drone scenario, which is the key context. When both modules are airborne, you eliminate ground reflections and multipath interference almost entirely, and that's where the link budget really opens up.
On the antenna side, the module is built around an external antenna connector (not a chip antenna), so you're not locked into a fixed radiation pattern. The "powerful RF front-end" language refers to the external amplifier stage and the receiver sensitivity optimization — that's the actual differentiator over commodity BLE modules that just run off the SoC's internal PA.
For the 50km drone-to-drone number, you'd want omnidirectional antennas with decent gain on both platforms — something like a 5-7 dBi dipole per side is a reasonable baseline assumption. You can push further with directional antennas but then you're trading tracking complexity for range, which only makes sense in fixed-link scenarios.
Honest answer though — for the exact connector spec and the antenna gain assumptions baked into the range figures, I'd drop a question directly to Moamen through the Dev Hub or pull the schematic from the GitHub repo (it's public at github.com/RFOXiA/MultiNav-Pro-Long-Range-BLE-Module-Firmware-STM32WB07). The RF front-end details tend to live in the hardware files more than the marketing copy.