I wait for this open-source model.

I should let the development agent select a model and run it to improve token efficiency.

Is it being used this much these days?