1h ago

Claude Code Cache Misses Cost 12.5x More Than Hits

——0——
Original post
OPMatt Henderson#914Matt Henderson@MATTHEN2

>You pay per token, in & out. Input tokens are cheaper if they’re cached, but it costs to write them to a cache, and the cache expires when we decide. If you add an image, it may invalidate the cache. We decide how many tokens to output, and most of them are hidden. Such a mess

5:32 AM · May 24, 2026 View on X
REPLYMatt Henderson#914Matt Henderson@MATTHEN2

Is there any other type of software with such complex billing?

A while ago, @RekaAILabs simplified all this by just charging per request, which seems so much more fair. Tokens aren’t universal anyway

Matt HendersonMatt Henderson@matthen2

>You pay per token, in & out. Input tokens are cheaper if they’re cached, but it costs to write them to a cache, and the cache expires when we decide. If you add an image, it may invalidate the cache. We decide how many tokens to output, and most of them are hidden. Such a mess

12:32 PM · May 24, 2026 · 1.4K Views
12:34 PM · May 24, 2026 · 593 Views
Claude Code Cache Misses Cost 12.5x More Than Hits · Digg