Applied AI·Fine-tuning and customisation
the frontier model proved the task works, so you train a small one on its outputs and cut the per-call cost by an order of magnitude.
Distillation
Draft summary, pending review
Fine-tuning a small model on a large model's outputs for one task, keeping most of the quality at a fraction of the inference cost. The standard cost-reduction path once a frontier model has proven the task in production. Check provider terms; some restrict training competitors on their outputs.