Is the solution to worsening usage limits using Luna for more things?

Like everyone else I want to use the smartest models for most of my tasks while staying efficient so right now I use 6.1 Sol and just change the reasoning. I'll have it spawn subagents of Luna for easy tasks. This works relatively well but I only sometimes use Luna agents. Maybe a solution is chatting with 6.1 Sol but have it always take actions by orchestrating a Luna subagent. I know context is limited with Luna but maybe with Sol orchestrating that won't matter as much.