LLM Cost & Token Economics: Pricing AI Features
Model what AI features really cost: tokens per request and per chat, cost per 1,000 requests, usage tiers, prompt caching and batch discounts, model routing, RAG context, cost per resolved ticket vs people, and gross margin by plan, on live auto-graded spreadsheets.
The first lesson is free — no account needed.
About this course
An LLM cost course that models what one fictional software company's AI features really cost, for product managers, founders, engineers and finance teams. You estimate tokens per request from word counts and see how a multi-turn chat multiplies the input tokens, look up input and output prices in a model price table for the cost per request and per 1,000 requests, build a monthly bill from light, regular and power users, price prompt caching and batch discounts and find the cache hit rate at which caching pays, route traffic between a large and a small model for the cheapest split that keeps quality, price a RAG feature from embeddings and retrieved context, compare the AI cost per resolved ticket with human cost and set usage limits that keep a 70% gross margin. Prices are illustrative tiers, not a vendor price list. The first lab is free.
Lessons
- Tokens per Request: Words, Context & Multi-Turn ChatsFree
- Cost per Request & per 1,000 Requests with a Price Table Pro
- Monthly Cost from Usage Tiers: Light, Regular & Power Users Pro
- Prompt Caching & Batch Discounts: When Do They Pay? Pro
- Model Routing: Blended Cost vs Quality Pro
- RAG Cost: Embeddings, Retrieved Context & the Monthly Bill Pro
- Cost per Resolved Ticket vs Human Cost: the Break-Even Pro
- Pricing an AI Feature: Gross Margin vs Usage Pro