How to choose from multi LLM models.

[blue-vision0]

I really appreciate your work in RoRF. However, I wonder why you just contrast two models in a single Random Forest Model.
So, if I want to choose from 6 LLM models and determine the best and worst for a specific prompt. Does it mean that I need to train 15 (C6,2) Random Forest Model to achieve that?
Or , I could just just modify the RoRF framework to train one Random Forest Model which can determine the probability of all that 6 models at once?