So you mean something like, "what if the baseline, off-the-cuff response for the next-gen models was tuned based on the results of the reasoning model excluding the reasoning itself?"

Exactly, albeit it may need the reasoning later to form the proper foundational logic in the weights.