Which do you prefer?
Difficult to say, considering what could be meant by 'traditional language'; the distinction made between the first and third options, vis a vis traditional language vs. 'high'-vernacular, seems to me to make sense mostly within the American context with its exceptional nature. In that arena, you would have English sparring off against Slavonic, for example. I still find the American situation too complicated to suggest an acceptable resolution, and so can only answer concerning the situation in the Levant.
In some way, Greek can be thought of as a traditional language in the Melchite/Antiochian Churches, and it still retains its place, though it has seen a lessening of its use with time in recent decades. It could and should certainly stay, I would think, but its use is marginal, and there doesn't exist friction between it and the Arabic language, which dominates the Levantine countries and the liturgical celebrations. Hence, it doesn't merit much study in the light of the topic of the thread.
In my Church's case, what pertains more to the choices presented by the question is the difference between modern Arabic vernacular and classical Fushah. In no framework would the former ever work properly. I find it interesting that you can notice across countries strong contrasts regarding the faceoff between 'old form' and present-day colloquial. Slavonic and Koine are not necessarily grasped or understood at all by Greeks and Slavs respectively, despite these being forms of their respective languages; in fact, I imagine Koine must seem to a Greek like Latin to a Spaniard or Italian, the languages of whom find their origins in Latin. I can see people from these countries arguing for modern Greek, Russian, and Italian services. Arabs on the other hand, have a unique privilege and noteworthy blessing. The classical form of their language enjoys the distinguished position of straddling the line between the realm of archaic languages that capture the august and rich qualities people look for in such lingual forms, and that of tongues people are quite familiar with at large. We have a strong familarity and grasp of the old Fushah, which sounds as distantly old as Elizabethan English in its distance from the colloquial. Regardless, it (or rather a newer form of it that incorporates more modern words into its vocabulary) is generally used when writing of reading, notwithstanding a few exceptions (there are some Robert Frosts amongst Lebanese poets who like to compose 'Lebanese' [ie. in Lebanese dialect] verses and poems). This general level of intimacy with Fushah allows us to fully understand the Liturgy whilst making a full departure from the informalities of vernacular dialects. The very suggestion that the Liturgy can be translated into 'Damascene' or the like is entirely inconceivable and nonsensical, unlike a Russian's preference for Russian over Slavonic (Slavonic, no question about it in my case).
1 (except for religious affiliation) and 4 seem right to me.
In IC XC