We present a highly unusual example of self-assembly, specifically a polar, mixed-ligand cage which forms in preference to symmetrical homo-ligand products, and which suggests that steric effects can be exploited to obtain novel non-uniform polyhedral cages. In particular, reaction between the bulky tripodal triphosphine 2,4,6-tris(diphenylphosphino)triazine, L1, the non-bulky tripodal trinitrile 2,4,6-tris(cyanomethyl)trimethylbenzene, L2 and silver hexafluoroantimonate, AgSbF6, in a 3:1:4 ratio gives the mixed-ligand aggregate [Ag4(L1)3(L2)(SbF6)]3+, 1-SbF6, instantly as the only product in quantitative yield. The X-ray crystal structure of complex 1-SbF6 is consistent with the suspected solution-state structure. The cage derives from trigonal-pyramidal geometry, the basal vertices of which are defined by three bulky triphosphines, L1, and the apical vertex by the non-bulky trinitrile, L2. There is apical elongation amounting to 19 % in comparison to the ideal uniform tetrahedron. The cage also encapsulates an SbF6 anion. 19F NMR spectra in solution for the analogous PF6 complex [Ag4(L1)3(L2)(PF6)]3+, 1-PF6, confirm that one anion is also encapsulated in solution. The synthesis of the analogous CF3SO3− complex, [Ag4(L1)3(L2)(OTf)]3+, 1-OTf, in solution is also described, although 1-PF6 and 1-OTf could not be isolated due to slow decomposition in solution. The selective formation of these mixed-ligand cages is discussed in terms of ligand–ligand and ligand-included anion steric repulsions, which we propose prevent the formation of the competing hypothetical homo-ligand tetrahedral structure [Ag4(L1)4(SbF6)]3+, and thus favour the mixed ligand cage. “Cage cone angles” for L1 and L2 are estimated at 115 ° and 101 °, respectively. Variable-temperature 31P NMR spectroscopy shows that complex 1-SbF6 and the related previously reported partial tetrahedral complex [Ag4(L1)3(anion)]3+ undergo dynamic twisting processes in solution between enantiomeric C3-symmetry conformations.