Claude Code

The Claude Limit Tier List: What Eats Your Usage, Ranked, With the Fix for Each

A coffee break, a compact and a team of Claudes all burn through your Claude limit in ways most people never see. Here's the full ranking from Claude Code's own cost docs, and the setting that fixes each one

Alex ChenAlex Chen@nocodealex · 50,000+ follow on Instagram
2min read
1prompts to copy
3numbered steps
8sections

Claude Code sends your whole conversation with every message. Most of the time that's cheap, because it's cached. But a few everyday habits make Claude pay for all of it again, or multiply it.

Here's the ranking, worst first, and the fix for each one.

S tier: a team of Claudes planning together

  • Agent teams use about 7x more tokens than a normal session when the teammates run in plan mode.
  • Each teammate has its own context window and loads your CLAUDE.md, MCP servers and skills on its own.
  • Fix: keep teams small, use Sonnet for teammates, keep spawn prompts short, and shut teammates down when their job is done.

A tier: the coffee break

  • Your first message after a break longer than the cache lifetime misses the cache and reprocesses your whole chat.
  • The lifetime is one hour on a subscription. It drops to five minutes on usage credits, and it's five minutes by default on an API key.
  • Fix: before a long break, finish the task and run /clear, or /rename the session so you can /resume it later.
  • Check it: /usage shows a Prompt cache line that says whether the cache is warm or cold, and how long you've been idle.

B tier: compacting a huge chat

  • /compact reads the whole conversation it summarizes, so compacting a large chat is itself a large request.
  • When you want a fresh start instead of continuity, /clear costs nothing.
  • Fix: /clear between unrelated tasks. When you do compact, tell it what to keep: /compact Focus on code changes and test output.

Free, for your email

Get the full setup steps + the copy-paste setup prompt

The rest of this guide (5 more sections, 1 copy-paste prompt) opens right here. The setup prompt, one paste into Claude that sets it all up, comes by email only. Use your real address.

50,000+ follow where these guides come from. No spam, unsubscribe in one click.

Read next

Set Up Claude Code Projects: One Goal, a Team of Claudes Working at Once
Projects turns one chat into a coordinator that splits your goal into threads, each its own Claude on its own branch, all sharing one memory. Who can use it now, how to set it up in five minutes, the first messages to send, and how to keep it from eating your usage
4 min
Claude Code's Effort Levels, Low to Ultracode: The Cheat Sheet for Which Level Fits Each Task
Claude Code has five effort levels and an ultracode switch, and the one you pick decides how fast you get an answer, how hard Claude checks its own work, and how soon you hit your usage limit. What each level really does, from Anthropic's own docs and tests, and a one-page cheat sheet for which one to use for each job.
5 min
Which Skill, Plugin or Server Is Eating Your Claude Limit: Read /usage
One command ranks skills, subagents, plugins and MCP servers by how much of your limit they use, and flags the habits behind it. How to read it, what to do about each culprit, and the catch
2 min