It depends on the accuracy you need.NMR is not good if you have many species and for amounts lower than 5%.The best is GC-MS: a first analysis will give you the species and an idea of the relative composition. Then if you want the exact composition you will have to calibrate the GC-MS with each molecule.

A side note on GCMS, there are some issues you can run into. Overlapping peaks (can be resolved using ion selection, i.e. SIM mose) and thermal degradation of oximes to benzonitriles (typically during injection, so injector temp is critical) just to name two.

There are multiple species in SX organics-the diluents are always multiple species taken in a specified boiling point range, some dearomaticised, some not. Various extractants also contain multiple isomers, for example LIX84, LIX984, Versatic 10. Others are more isomerically pure, like TBP, DEHPA and Cyanex 272, but they can also contain impurities from the synthesis. So your SX organic could easily contain over 200 discrete species in solution. The only way to try to resolve them is with a chromatographic technique such as GC-MS, and not an 'all in one' analysis technique like NMR or FT-IR. The suggestion regarding ion selective mode on a GC-MS is a good one if you are really wanting to chase down every single peak you can see in the system, but I feel you may not really want that, and the collectives of 'diluent peaks' and 'extractant peaks' etc might be more appropriate to your needs. So my question to you is "what are you trying to find out?" Are you interested in reagent degradation products? Impurity ingress? If you have a species adversely affecting phase separation for example, it could be one that you cannot see using GC-MS, even after derivatisation. It can be real 'needle in a haystack' stuff.

It depends on your practice.

If you want to analyze the chemical elements, you will use the MS or NMR.
But if you only want to analyze the material composition, you just need to use HPLC, or GC, no need to use GC-MS.