Claude Prompt Caching Minimum: 1,240 Haiku 4.5 Calls, 0 Cache Hits
I moved a scoring job from Sonnet 4.5 to Haiku 4.5 to cut its cost by two thirds. A week later the bill had dropped, but nowhere near as much as my spreadsheet promised. The cause was sitting in every response I'd logged: cache_read_input_tokens: 0 . All 1,240 calls. No error. No warning. The cache…
Read the full story at DEV Community — AI ↗
Timeline · 1 report
- 2026-11-12 19:27 · DEV Community — AI
Claude Prompt Caching Minimum: 1,240 Haiku 4.5 Calls, 0 Cache Hits