Curated references that pair with the system-prompts archive—starting with agent routing and model selection.
Tiered LLM model stack: costs, context, benchmarks, and when to use each model—responsive cards on mobile, full table on larger screens.