PriceIndex / Editorial

Prompt Caching articles & guides

Caching repeated context.

Prompt Caching

Abstract technical illustration of prompt cache memory blocks routing LLM requests between cached reads and cache writes

Prompt Caching / Analysis

Prompt Caching Economics: Break-Even Math for OpenAI vs Anthropic

Anthropic's prompt caching charges an upfront write multiplier that penalizes low hit rates, whereas OpenAI's legacy prefix caching carries zero write surcharge. Here is the mathematical derivation of exact break-even hit rates and cost trade-offs across 1k, 10k, and 100k token prompts.

Oct 5, 202610 min read