docsfoundations2 min read

When You Run Out of Usage in Claude Code

Hit your 5-hour or weekly limit? What each message means, why switching models usually won't help, how auto-continue works, and when usage credits are worth turning on.

Your plan has two clocks

On Pro, Max, Ultra, Team and Enterprise, Claude Code draws from an allowance that resets on two windows: a rolling five-hour window and a weekly one. Claude chat and Cowork draw from the same allowance. When either window runs out, Claude stops mid-task and tells you which one.

Read the message first

MessageWhat it meansWhat helps
"You've hit your 5-hour limit" or "weekly limit"Your plan allowance for that window is used up. It covers every model.Wait for the reset, or turn on usage credits. Switching models won't help.
"You've hit your Opus limit" or "Sonnet limit"You've used up one model family's share.Switch to a model outside that family with /model and keep going.
"You've hit your individual spend limit" or "org's monthly spend limit"You're on usage credits and they hit a cap.Raise the cap, or wait for the plan reset if the message names one.
A context or auto-compact warningNot a usage limit. The conversation is nearly full.See context window.

Claude Code renamed the "session limit" to the "5-hour limit" in 2.1.312. Same window, clearer name. Source for the table: Anthropic's manage costs page, read 16 November 2026.

Let it wait for you

Since 2.1.234, Claude Code waits in the open session and picks the task back up after the reset. It's on by default in interactive sessions signed in with a claude.ai subscription. You'll see two lines at the bottom:

text
Usage limit reached · limit resets 3:45pm
Continuing automatically at 3:45pm · Esc to cancel

Keep the session open. At the reset, Claude continues where it stopped; it doesn't resend your last message. If the laptop slept through the reset for more than about 30 minutes, press Enter to continue.

Three things to know:

  • Permission prompts still stop it. The continued task asks for permissions as usual, so it can sit on a prompt while you're away. Why unattended runs stop covers how to avoid that.
  • It gives up after three tries. After three automatic continuations, run /rate-limit-options to choose what happens next.
  • Esc cancels it. So does /rate-limit-options and picking Don't continue automatically.

The options menu

Since 2.1.323, when a model needs usage credits that are off or used up, Claude Code shows a menu: stop, switch to usage credits, or upgrade. Fast mode always needs credits, so this is where most people first see it.

Usage credits: pay to keep going

Usage credits let you work past your plan's limit at API prices. Run /usage-credits:

  • On Pro, Max or Ultra it opens Settings > Usage on claude.ai, where you turn credits on and set a monthly spend limit.
  • On Team or Enterprise with billing access it opens your organization's usage settings.
  • On Team or Enterprise without billing access it asks you to confirm, then sends a request to your admins.

Set a monthly spend limit when you turn them on. Without one, /usage shows "Unlimited".

Hit the limit less often

Run /usage. On a subscription it shows what used your allowance (skills, subagents, plugins, MCP servers) and suggests changes for anything behind 10% or more of recent usage. Press f to have Claude Code apply the ones it can.

The usual culprits are long sessions that never get cleared, a scheduled task firing every few minutes, and effort left at max. Cost optimization and prompt caching go deeper.

New guides, when they ship

One email, roughly weekly. CLAUDE.md templates, workflows I actually use, and the cut-for-length stuff that does not make the public guides. One-click unsubscribe.

Subscribe on Substack

Opens Product Field Notes, my Substack, in a new tab. You enter your email there.