The systems guide to production token optimization

The New Stack · 1 month ago · 24

Enterprise AI support and CI agent systems (Concierge and Pathfinder) saw rapidly growing latency and costs as autoregressive history re-billing compounded across multi-step runs. The article’s baseline pricing example is $3 per 1M input tokens and $15 per 1M output tokens. It proposes fixes including dynamic context injection/RAG, prompt compression (LLMLingua-2), strict JSON/schema enforcement, output token bounding, caching with up to 90% discounted reads, semantic caching, context compaction, and model cascading to keep token growth bounded.