Just specify that the more efficient model should only be used if the cost of doing the task would be less than using the higher horsepower model. Fable has no problem routing based on that instruction.
The timing (after 5pm ET on a Friday) is telling. Build a KYC module over the weekend and we’ll be back on Fable after uploading our ID Monday morning.
I'm from Europe, but I think this move can actually achieve America First, at least as a first-order effect. The model is incredibly strong; I've used it for a few days. It gives anybody with access a serious boost. If they also take Opus and GPT5.5 away from me it's going to really be a drop in productivity.
The difference between "being a snowflake" and "having a point of view" revolves around who's talking to me and whether or not they want something. If comparing yourself to others is a slow form of suicide, letting people make that comparison for you is madness.