Guide · 6 November 2026 · 6 min read

Claude Code usage limit: how to check what's left

Your Claude Code usage limit is not a number of messages: it is a share of your Claude plan that fills up as you work and resets on a schedule. Most paid plans have two windows, a session that resets every five hours and a weekly limit, and Claude Code draws from the same pool as the Claude app. Here is where to read each one, what to do when you hit it, and how the Codex limits work if you use both agents.

Check your usage inside Claude Code

Type this in any session:

/usage

The screen shows your session cost, your plan usage limits and your activity stats. On Pro, Max, Ultra, Team and Enterprise plans it also includes a breakdown of what counts against those limits. /usage runs immediately, even while Claude is in the middle of a response, so you do not have to wait for a turn to end.

Do not confuse it with /status: in Claude Code, /status shows the version, the model, the account and the connection, not your limits.

Check it on claude.ai

In the Claude app or on claude.ai, open Settings → Usage. Anthropic’s help centre is explicit that your combined usage across the Claude app and Claude Code counts toward your limits. A long chat in the browser in the morning leaves less room for Claude Code in the afternoon.

How the limits are built

According to Anthropic’s help centre, on the paid plans (Pro, Max and Ultra):

• 5-hour session limit. Most paid plans have a session limit that resets every five hours. Verified Ultra accounts do not have one.
• Weekly limit. It resets once a week.
• Per-model limit. Max and Ultra plans also have a separate weekly limit for Fable.

Anthropic does not publish a fixed number of prompts per window: how fast you use it depends on the model, the length of the conversation and the files Claude reads. Treat figures you find on third-party sites as estimates.

What to do when you reach the limit

1. Read the reset time in /usage. If the 5-hour window resets in twenty minutes, waiting is often the cheapest option.
2. Check whether it is the session or the week. A full 5-hour window comes back the same day; a full weekly limit does not.
3. Add usage credits if your plan offers it: /usage-credits opens the billing settings (on Team and Enterprise, it sends a request to your admin).
4. Switch agent. If you also have a ChatGPT plan, Codex has its own limits, separate from Claude’s.

Habits that make the limit last longer

• One task per session. Every request carries the conversation so far, so a long history makes each new message heavier. Start a new session for an unrelated task.
• Mind long resumes. When you resume a long conversation after a long break, Claude Code may show a “Resume this conversation?” dialog with the share of your 5-hour limit the saved conversation alone would use. If the next task does not need that history, start a new conversation instead.
• Be careful with parallel sessions. Several sessions running at once draw on the same limits. See how to run multiple Claude Code sessions at once without losing track of them.

Codex usage limits, for comparison

Codex, OpenAI’s coding agent, is included in ChatGPT plans (Free, Go, Plus, Pro, Business, Edu and Enterprise, according to OpenAI’s pricing page). Its limits follow a similar logic:

• Plus and Standard Business plans have limits per five-hour period; Pro plans currently have no five-hour limit.
• Weekly limits may also apply.
• During a Codex CLI session, type /status to see your remaining limits.
• The dashboard at chatgpt.com/codex/settings/usage shows current limits and reset times.

The two plans are independent: reaching your Claude limit does not touch your ChatGPT one, which is why many developers keep both agents within reach. More on that in Claude Code and Codex together.

Keep both limits in view with Tarmac

/usage and /status answer the question when you ask it. Tarmac, a free and open source local dashboard for macOS and Windows (WSL 2), keeps the answer on screen:

• Claude usage in the sidebar: your plan (Pro, Max, Team…), the 5-hour session, the week and each per-model weekly limit, with the percentage used and the reset time. Gauges turn orange from 70 % and red from 90 %.
• Codex usage right under it: your ChatGPT plan, the 5-hour and weekly windows and when they reset.
• How it reads them. For Claude, Tarmac uses Claude Code’s own sign-in, reads it from where Claude Code keeps it, and sends it only to Anthropic’s usage endpoint, at most once a minute and only while a Tarmac window is visible. For Codex, it reads the limits Codex already writes in its local session files: no token and no network call.
• The sessions next to the gauges. Every Claude Code and Codex session of every project, so you can see which one is using your plan.

Both panels can be turned off in Settings. The details are in the Claude Code and Codex page.

Next steps

Install Tarmac for free and open the sidebar: your Claude plan appears as soon as Claude Code is signed in, and your Codex plan after your first Codex exchange on that machine.

Sources: Claude Code commands, Manage sessions, Manage usage credits for paid Claude plans and Codex pricing, checked on 6 November 2026. Plans and limits change: the official pages are the reference.