If we could, we'd all be running Astra + Fable 5.1 all day
But we can't afford it. So > 50% of our inference is running on other models. More as we add cheap labelling options like Jev.
We all want to be running 2-10 agents on Astra + Fable 5.1
We just can't afford it