Hit your Claude Code limit? What to do next
The message says which limit you hit and when it resets. Below is what each one means and how to keep working: wait, pay, switch models or switch agents.
Short answer. For a session or weekly limit, switching models won’t help: wait for the reset shown in the message, turn on usage credits with /usage-credits, or continue the task in another agent such as Codex or OpenCode. For an Opus or Sonnet limit, /model to another model family keeps you going.
Which limit did you hit?
| Message | What it is | Does /model help? |
|---|---|---|
You’ve hit your session limit · resets 3:45pm | The rolling five-hour window | No, it covers all models |
You’ve hit your weekly limit · resets Mon 12:00am | The weekly limit on Pro and Max | No, it covers all models |
You’ve hit your Opus limit · resets 3:45pm | A limit on one model family (Opus or Sonnet) | Yes, pick a model outside that family |
Before a limit runs out, Claude Code warns you with a line like You’ve used 85% of your session limit · resets 3:45pm. Run /usage any time to see both bars and the reset times.
Keep working right now
1. Let Claude Code wait and continue
Since v2.1.234, Claude Code waits in the open session and continues the interrupted task on its own after the reset. This is on by default and Esc cancels it; /rate-limit-options lets you choose between waiting, adding credits and upgrading. In the Claude desktop app, the session-limit card has an Auto-continue when limits reset checkbox. Good for a long task you were going to leave running anyway.
2. Pay for the overflow with usage credits
On Pro and Max, /usage-credits opens Settings → Usage on claude.ai, where you can turn usage credits on and set a monthly spend limit. Past your plan’s limit you then pay standard API rates. Note that the prompt cache lifetime drops from an hour to five minutes while you draw on credits, so long pauses cost more.
3. Use a free limit reset, if you have one
Anthropic occasionally gives eligible plans a one-time limit reset that refills the session or weekly limit at once. It lives in Settings → Usage on claude.ai and can’t be used from the terminal.
4. Switch to another model, if the limit is model-specific
After “You’ve hit your Opus limit”, /model to Sonnet (or the other way round) keeps the same session going.
5. Continue in Codex or OpenCode
If you also pay for ChatGPT or OpenCode Go, the same repository is one command away: codex or opencode. The hard part is context, so make the hand-off cheap before you need it:
- Commit small and often, so the other agent starts from a clean tree.
- Keep project instructions where every agent reads them. Claude Code reads
CLAUDE.md; Codex and OpenCode readAGENTS.md. A one-lineCLAUDE.mdthat points toAGENTS.mdkeeps them in sync. - For long tasks, ask Claude to keep a short progress note (what’s done, what’s next) in the repo while there is still room in the window.
This only works if you know the other agent has room. Codex and OpenCode have five-hour and weekly windows of their own; here is how they compare.
Make the next five hours last longer
Claude Code sends the whole conversation with every request, so the size of your context matters more than the number of messages. From Anthropic’s own advice:
/clearbetween unrelated tasks. A one-line question in a session open all day still pays for the whole history.- Match the model to the job. Sonnet handles most coding; keep Opus for hard reasoning. Lower
/effortfor simple tasks. - Turn off MCP servers you don’t use in
/mcp, and check/contextfor what is taking space. - Mind breaks. The first message after more than an hour away misses the cache and reprocesses everything.
- Watch background work. Scheduled loops, subagents and agent teams all draw on the same limits, even while you are idle.
- Read the breakdown. On Pro and Max,
/usageshows which skills, subagents and MCP servers used the most, and flags habits like long context.
See it coming instead of hitting it
Claude Code warns at 85%, but only in the session you are typing in. If you work across the terminal, an IDE and claude.ai, or across Claude Code and Codex, the warning tends to come too late to plan around.
CapaTheNotch: limits you can see before you hit them
The MacBook notch shows how much is left in the five-hour and weekly windows of Claude Code, Codex or OpenCode Go, and when each resets, so you know which agent has room before you start a task. One notification when a window drops below 10%. Free and open source, for Apple Silicon Macs on macOS 14 or later.
Download for macOSSources
- Error reference (Claude Code docs)
- Manage costs effectively (Claude Code docs)
- What is a limit reset? (Anthropic)
- Claude plans and pricing (Anthropic)
FAQ
Does switching models help when I hit the Claude Code limit?
Only for a model-specific limit such as “You’ve hit your Opus limit”. The session and weekly limits are shared across all models, so /model won’t get you past them.
Can Claude Code continue by itself after the reset?
Yes. Since v2.1.234 Claude Code waits for the reset and continues the interrupted task on its own; this is on by default. In the desktop app, the session-limit card has an “Auto-continue when limits reset” checkbox.
Do claude.ai chats use up my Claude Code limit?
Yes. On Pro and Max, Claude and Claude Code share one set of usage limits.
How long do I have to wait?
The message itself says when the limit resets, for example “resets 3:45pm”. A session limit is at most five hours away; a weekly limit resets at the fixed weekly time assigned to your account.