It's Wednesday, you're in the middle of a refactor, and Claude Code says "You've hit your weekly limit · resets Mon 12:00am." The five-hour window you were watching looked fine. This guide explains how the limits actually work on Pro and Max, every way to see where you stand before it happens, what your options are when it does, and how Codex's limits compare.
Quick answer
- Two limits at once. Pro and Max have a 5-hour session window and a weekly limit. Hitting either one stops you. Max also has a separate weekly limit for Fable models.
Shared with chat. The same allowance covers Claude on the web, desktop, mobile and Claude Code.
Check it in Claude Code with
/usage. It shows your plan's bars and reset times. On the web: claude.ai → Settings → Usage.Always visible: the status line can show
5h: 23% 7d: 41%from the officialrate_limitsfields.Out of quota? Wait for the reset, upgrade, or turn on usage credits (billed at API rates). Switching models doesn't help with the session or weekly limit.
- Codex: check with
/statusin the CLI or at chatgpt.com/codex/settings/usage.
How the two limits work
The 5-hour window. Your usage is measured over a rolling five-hour session. When it fills up, you wait until it resets. Anthropic doesn't publish a fixed number of messages, because it depends on the model, the length of the conversation and how much context each request carries. In May 2026 Anthropic doubled Claude Code's five-hour limits for Pro, Max, Team and seat-based Enterprise.
The weekly limit. On top of the window there's a weekly cap. It resets at a fixed time each week assigned to your account (you can see it in Settings → Usage). It's not a rolling seven days. Weekly limits started in August 2025. This is the one that surprises people: you can have plenty of room in the current five-hour window and still be blocked for days.
Per plan:
| Plan | 5-hour window | Weekly limit | Notes |
|---|---|---|---|
| Pro | Yes | Yes | The base amount |
| Max 5x | Yes | Yes, plus a separate Fable weekly limit | 5× Pro's usage |
| Max 20x | Yes | Yes, plus a separate Fable weekly limit | 20× Pro's usage |
Chat counts too. Long conversations on claude.ai come out of the same allowance as Claude Code.
Every way to check your usage
/usage in Claude Code
Type /usage (/cost and /stats are aliases). For Pro and Max it shows:
- bars for the current session and the week, with reset times;
- activity stats, with
d/wfor the last 24 hours or 7 days; - a breakdown of what used the most: skills, subagents, plugins, MCP servers, long context, cache misses.
Two things to know:
- The breakdown comes from the history on this machine. Usage from claude.ai or another computer counts against your limit but won't show up in the breakdown.
- The dollar figure is an estimate at API list prices. On a subscription you're not billed that. It's useful as "how heavy was today", not as a bill.
/context shows how full the current context window is, which is a different thing from your plan limit.
The status line (always on screen)
Claude Code passes your limits to the status line script: rate_limits.five_hour.used_percentage, rate_limits.seven_day.used_percentage and their resets_at times. Run /statusline and ask for something like 5h: 23% 7d: 41%. The docs include a ready-made example. These fields appear for Pro and Max subscribers after the first response of a session. Community status lines such as >ccstatusline do this too.
claude.ai → Settings → Usage
The web page (>claude.ai/settings/usage) shows the current session with time remaining, the weekly limits with their reset, and your usage credits. It's the only view that includes every device and the chat apps.
Warnings and messages
Claude Code warns you before you run out, with messages like "You've used 85% of your session limit · resets 3:45pm." When you hit a limit you'll see one of these:
You've hit your session limit · resets 3:45pmYou've hit your weekly limit · resets Mon 12:00amYou've hit your Opus limit/You've hit your Sonnet limit: these are per-model-family limits. Switching to a model outside that family keeps you going.
What to do when you hit the limit
- Let it finish the step. If the five-hour limit lands mid-response, Claude Code may wrap up the current step first ("Usage limit reached · wrapping up").
- Let it continue by itself. Recent versions can wait in the open session and resume after the reset. It's on by default for subscribers (
/config→ "Continue automatically at usage limit"). It won't wait for a reset more than 24 hours away, so it doesn't help with the weekly limit. - Usage credits. Turn them on in Settings → Usage (or
/usage-creditsin Claude Code). Usage past your limits is billed at API rates, separately from the plan, with an optional monthly cap. - Upgrade from Pro to Max 5x, or from 5x to 20x.
- Use API credits from the Claude Console instead of the subscription.
Changing models with /model does not restore access when the session or weekly limit is the one you hit. Those are shared across all models.
Make the week last
All of these come from Anthropic's own cost guidance:
/clearbetween unrelated tasks. The whole conversation is re-sent on every request, so old context makes every later message cost more./renamethe session first if you want to/resumeit later./compactat natural breaks, with a focus:/compact Focus on the API changes.- Pick the model at the start. Sonnet handles most coding work for less than Opus. Changing models mid-session drops the prompt cache, and the next turn re-reads everything at full cost.
- Lower the effort with
/effortfor routine work. Thinking tokens are billed as output. - Keep CLAUDE.md short (the docs suggest under 200 lines) and turn off MCP servers you don't use (
/mcp). - Long breaks are expensive. On a subscription the cache lasts about an hour. Come back after lunch to a huge session and the first turn re-reads it all uncached.
- Go easy on big screenshots. Lots of full-screen images fill the request and push Claude Code to drop old ones, which means slower, costlier turns. Send the text of what you mean, or a small crop, when you can.
- Parallel sessions multiply usage. Five agents at once spend the week five times faster. That's fine as long as you can see it happening.
Codex limits
Codex limits depend on your ChatGPT plan:
- Plus and Business: a five-hour window, measured in local messages (OpenAI publishes ranges per model), and weekly limits may also apply.
- Pro: currently no five-hour limit. Weekly limits aren't published.
- Fast modes cost more: fast mode uses 2.5× the included usage and ultrafast 8×.
Check where you stand with /status inside the Codex CLI, or on the dashboard at >chatgpt.com/codex/settings/usage. /usage daily and /usage weekly show your token use over time.
Tools that track it for you
- >ccusage:
npx ccusage@latest. Daily, weekly, per-session and per-5-hour-block reports for Claude Code, Codex and other CLIs, computed from the logs on your machine (~/.claude/projects/,~/.codex/sessions/). - >Claude-Code-Usage-Monitor: a live terminal monitor with forecasts (
claude-monitor). - >CodexBar: a macOS menu bar app with session and weekly limits for Claude, Codex and other providers.
Where Aki fits
Aki is a free macOS app. Among other things, it puts your week on the edge of the screen:
- One ring for Claude, one for Codex, showing how much of the weekly limit is used. They turn yellow at 50%, orange at 70% and red at 90%. A second figure shows today's share of the week.
- Hover for the details: the 5-hour window, when each limit renews, whether you're ahead of or behind the pace, today's tokens and what they'd cost at API prices, the last 7 days and the projects that used the most.
aki usageprints the same in the terminal.- "Only my week" mode asks for no permissions at all. Aki reads what Claude Code and Codex already leave on your Mac (the status line report and the session logs) and never asks a server.
Aki's main job is something else. You point at things on screen (an element in Chrome, an area, a terminal error) and it sends them to the right agent session. You can use either part, or both. It's Mac-only (Apple Silicon, macOS 14+), and there's no account.
FAQ
What is the Claude Code weekly limit?
It's a cap on how much you can use Claude per week on Pro and Max, on top of the 5-hour session window. It resets at a fixed time each week assigned to your account. Max plans also have a separate weekly limit for Fable models.
How do I check my Claude Code usage?
Type /usage in Claude Code to see your session and weekly bars with reset times. You can also open claude.ai → Settings → Usage, or add the limits to your status line.
When does the Claude Code usage limit reset?
The session limit resets on a rolling five-hour window. The weekly limit resets at a fixed weekly time shown in Settings → Usage and in the limit message itself (for example, "resets Mon 12:00am").
Does switching models help when I hit the limit?
Not for the session or weekly limit, because those are shared across all models. It only helps with a model-family limit ("You've hit your Opus limit"). Then you can switch to a model from another family.
Do Claude chat and Claude Code share the same limit?
Yes. Usage on claude.ai (web, desktop, mobile) and in Claude Code comes out of the same allowance.
How do I check my Codex usage limit?
Run /status in the Codex CLI, or open chatgpt.com/codex/settings/usage. Plus and Business have a five-hour window. Pro currently has none.
Sources
- https://support.claude.com/en/articles/11145838-use-claude-code-with-your-pro-or-max-plan
- https://support.claude.com/en/articles/11647753-how-do-usage-and-length-limits-work
- https://support.claude.com/en/articles/8325606-what-is-the-pro-plan
- https://support.claude.com/en/articles/9797557-usage-limit-best-practices
- https://support.claude.com/en/articles/12429409-manage-usage-credits-for-paid-claude-plans
- https://claude.com/pricing
- https://www.anthropic.com/news/higher-limits-spacex
- https://code.claude.com/docs/en/costs
- https://code.claude.com/docs/en/commands
- https://code.claude.com/docs/en/errors
- https://code.claude.com/docs/en/interactive-mode
- https://code.claude.com/docs/en/statusline
- https://code.claude.com/docs/en/prompt-caching
- https://code.claude.com/docs/en/claude-directory
- https://techcrunch.com/2025/07/28/anthropic-unveils-new-rate-limits-to-curb-claude-code-power-users/
- https://learn.chatgpt.com/docs/pricing
- https://learn.chatgpt.com/docs/cli/slash-commands
- https://github.com/ryoppippi/ccusage
- https://github.com/Maciek-roboblog/Claude-Code-Usage-Monitor
- https://github.com/steipete/CodexBar
- https://useaki.vercel.app