Batch APIs & Distillation for Cost

Cut LLM costs with two offline levers — batch APIs that trade latency for 50% off, and distillation into a cheap model that mimics the frontier — with when to use each and the full cost playbook.