Model Selection and Optimisation

Pick models, control cost and latency, use caching and understand tokens and pricing.

1. 5.1Choosing a ModelHow the Claude model tiers differ, how to read model IDs and aliases, and a repeatable process for picking and upgrading models without surprises.8 min
2. 5.2Tokens, Pricing and CountingWhat a token is, how input and output tokens are priced differently, how to count them before you send a request, and how to turn the usage object into a cost figure.8 min
3. 5.3Prompt Caching for DevelopersHow prompt caching works as a prefix match, where to place cache_control breakpoints, what it costs, and how to prove from the usage object that it is working.10 min
4. 5.4Latency, Throughput and Rate LimitsWhat drives response time, how rate limits are defined and enforced, which errors mean retry versus fix, and the techniques that keep a Claude integration fast under load.9 min