How to Reduce LLM API Costs: A Practical Checklist for Caching, Batching, and Model Routing

Five levers that cut a real LLM API bill: caching, batching, model routing, prompt compression, and output limits, with a worked Novita AI cost example.