Mo and its alloys have been widely used as refractory materials owing to their excellent high temperature properties, but a critical limitation is their low ductility. Doping the grain boundaries (GBs) of Mo with metals such as Zr or Al have previously been demonstrated as a promising approach to address this shortcoming, whereas other alloy elements are known to embrittle the GBs. In this work, we investigated the segregation and strengthening/embrittling effects of 29 metallic dopants at the Σ5 (310) tilt and Σ5 (100) twist Mo GBs using density functional theory (DFT) calculations and empirical continuum models. In agreement with previous works for other metals, we find that the strain, as measured by the relative metallic radius versus Mo, is a good predictor of the segregation tendency, while the difference in cohesive energies between the dopant and Mo is a good predictor of the strengthening/embrittling effect. However, we find that dopant chemistry also plays a significant role in affecting segregation behavior at GBs, particularly in driving the formation of intermetallic precipitates or 2-D interfacial phases (complexions). We also show that the site preference of a dopant in the GB can lead to strengthening effects that deviate from those predicted using simple bond-breaking arguments. Assuming a fast cleavage model of fracture, Ta, Re, Os and W are predicted to have a weak strengthening effect on Mo for the Σ5(310) tilt GB, and Mn, Fe, Co and Nb are predicted to have reasonable strengthening effects for the Σ5(100) twist GB.