What is Claude Code token usage breakdown?

Typically 97% cache reads in heavy sessions, per this 187-session analysis. Controllable output tiny at 6.6%.

How to reduce Claude Code token burn?

Use /compact early, grep over agents, avoid peaks. Tools like ccwhy reveal leaks.

Is ccwhy safe for Claude data?

Yes – offline Rust CLI on local ~/.claude/. No API calls.

187 Claude Code Sessions Burned $6,744 in Tokens – 97% Was Pointless Cache Reads

Everyone thought Claude Code was the efficient AI coder's dream. Turns out, it's gobbling tokens like a Silicon Valley VC at a free buffet – mostly on re-reading its own context.

DevTools Feed Apr 03, 2026 3 min read 23 views

Read in: Deutsch English Español Français Italiano 日本語 한국어 Português (BR) Русский Türkçe

Pie chart breaking down 3.3 billion Claude Code tokens: 97% cache reads dominating

⚡ Key Takeaways

97% of Claude Code tokens are cache reads, re-reading context every turn. 𝕏
Simple tweaks like /compact and grep slash burn by avoiding bloat. 𝕏
ccwhy tool open-sourced: Analyzes your sessions offline to pinpoint waste. 𝕏

Published by

DevTools Feed

Ship faster. Build smarter.

#AI coding costs #Anthropic quota #claude-code #token usage

Worth sharing?

Get the best Developer Tools stories of the week in your inbox — no noise, no spam.

Originally reported by dev.to

⚡ Key Takeaways

The 60-Second TL;DR

DevTools Feed

Share this article

Worth sharing?

Related Stories

Burned $6,744 on Claude Code Sessions—97% Went to Cache Reads, Not Code

Claude Code Turned My Empty Folder into a Portfolio—But Tokens Tell the Real Story

Claude Code's Three Layers: Hooks, MCP, Skills Dissected

Claude Code Limits? Downgrade and Compact Now

Stay in the loop