How to check your Claude Code and Codex usage limits in 2026: 8 ways compared

Disclosure: Enjoy is our product, and it shows your agents’ usage limits. We checked every way below against Anthropic’s and OpenAI’s documentation and each maker’s own pages on November 2, 2026, and we say where another way is the better pick, including the free ones built into Claude Code and Codex.

The ways to check Claude Code and Codex usage, compared

Both labs ration their plans over two windows. Anthropic’s paid plans have “a session limit that resets every five hours, and weekly limits that reset once every week,” and usage across claude.ai, Claude Code and Claude Desktop “counts towards the same usage limit.” On ChatGPT Plus, OpenAI warns that “You may reach the five-hour limit before five hours have passed, even if you still have weekly usage remaining.” Claude’s verified Ultra accounts and ChatGPT’s Pro plans have no five-hour limit. And since mid-October, Claude subscribers have reported their week running out sooner for the same work. Every way below shows what’s left, or estimates it. They differ in where the numbers appear, whether they stay on screen, and how many accounts and agents they cover.

Way	Best for	Price, checked November 2, 2026	Runs on	Agents
/usage and Settings > Usage	Checking Claude now and then	Included in Claude Pro ($20 a month), Max ($100), Ultra ($200), Team and Enterprise	Claude Code, its VS Code extension, claude.ai	Claude Code
/status and Settings → Usage	Checking Codex now and then	Included in ChatGPT plans; the CLI from Plus ($20 a month)	The Codex CLI and IDE extension, ChatGPT	Codex
Your status line	Your limits always in view in a terminal	Free with Claude Code or Codex	Claude Code’s and Codex’s terminal apps	Claude Code, Codex
CodexBar	A menu bar meter for both, and most other plans	Free, open source	Mac; Linux desktop app	Codex, Claude and 109 more providers, by its count
Usagebar	Claude Code alerts on a Mac	$9 once	Mac	Claude Code
ccusage	Tokens, and what they’d cost at API prices	Free, open source	A terminal, through npx or bunx	Claude Code, Codex and 16 more
Orca	Every account’s limits beside your agents, free	Free, open source	Mac, Windows, Linux; phone companion	Claude Code, Codex, Grok, Antigravity and 6 more
Enjoy	Every account’s limits, and moving on when one runs out	A monthly subscription, free for 7 days, with your phone and other computers (pricing)	Mac, Windows, Linux	Claude Code, Codex, Grok Build, Antigravity, Muse Code

Prices are as each maker lists them, before tax. Every way reads the plan you’re signed in to, and Claude’s pricing lists Claude Code on Pro, Max and Ultra, not Free. Codex’s pricing gives Free, at $0, and Go, at $8 a month, Codex in the desktop app, and Pro starts at $100.

Claude Code’s own: /usage and Settings > Usage

Best for: checking your Claude limits now and then, at no cost.

Run /usage in Claude Code. Anthropic’s cost docs say “Subscribers see plan usage bars, activity stats, and a usage breakdown on the same screen,” and on a Pro, Max, Ultra, Team or Enterprise plan the breakdown shows “what counts against your plan limits, with recommendations”: the share of recent usage that came from skills, subagents, plugins and each MCP server, and a suggested change for anything behind 10% or more of it. /cost and /stats open the same screen, and Claude Code runs /usage “immediately without interrupting the response,” per its commands reference, so you can check in the middle of a task. In the VS Code extension, /usage opens “the Account & usage dialog,” with “usage bars for your plan’s limits, such as the current session and the week.”

Away from the terminal, open Settings > Usage on claude.ai. Anthropic’s help center says it shows “progress bars showing how much of your five-hour session and weekly usage limits you’ve consumed,” the time left in the current session, and when your weekly limits reset, “for all models, and for Fable (if included in your plan).” Its limits article says weekly limits “reset at a fixed time each week that is assigned to your account,” and on Ultra, “verified accounts won’t see a five-hour session limit.”

Price: included in every plan with Claude Code: Pro at $20 a month, or $17 a month billed yearly, Max at $100, Ultra at $200, Team and Enterprise.

Watch for: you have to go and look. The breakdown in /usage is “approximate and computed from local session history on this machine, so usage from other devices or claude.ai is not included,” though the bars above it come from Anthropic. When Anthropic’s usage endpoint is busy, /usage shows “the last usage bars it loaded on this machine within the past 60 minutes,” and a message such as “rate limited until 3:14am” “means Claude Code can’t fetch your plan limits again until then, not that your work with Claude is limited.” With an API key there are no plan bars, only the session’s own cost and tokens.

Codex’s own: /status and Settings → Usage

Best for: checking your Codex limits now and then, at no cost.

In the Codex CLI, run /status. Codex’s pricing page says “If you want to see your remaining limits during an active Codex CLI session, you can use /status,” and in the IDE extension, /status shows “the chat ID, context usage, and rate limits,” per Codex’s command docs. Everywhere else, OpenAI points to one page: “Check Settings → Usage for your remaining allowance and reset times,” says its help center, in ChatGPT or the ChatGPT desktop app, and the pricing page adds that “You can find your current limits in the usage dashboard.” The CLI’s /usage is about tokens rather than limits: it shows “daily, weekly, or cumulative ChatGPT token activity,” and redeems an earned reset when you have one.

On Plus, the five-hour and weekly limits apply together, and “you need allowance remaining in both to continue.” Pro 100, Pro 200 and Pro 500 “currently have no five-hour usage limit in Work and Codex.”

Price: free with your ChatGPT plan. Codex’s pricing lists the CLI and the IDE extension from Plus, at $20 a month, and Pro at $100, $200 or $500.

Watch for: /status prints a summary inside a running session; for a reading that stays on screen, use the status line below. And the pool is about to change: “Starting in December, Plus, Pro, and Business plans’ usage allowance will be shared across ChatGPT and Codex,” says OpenAI’s usage article, and “You will be able to see your combined remaining usage and reset times in Settings → Usage.” Our memo explains what changes.

Your status line, in Claude Code or Codex

Best for: keeping your limits in view while you work in a terminal, free.

Claude Code’s status line runs a script of yours and shows what it prints, and the data it hands that script includes your limits: rate_limits.five_hour.used_percentage and rate_limits.seven_day.used_percentage, the “Percentage of the 5-hour or 7-day rate limit consumed, from 0 to 100,” each with a resets_at time. You don’t have to write the script. “The /statusline command accepts natural language instructions describing what you want displayed,” so you can ask it for your five-hour and weekly usage, and Claude Code “generates a script file in ~/.claude/ and updates your settings automatically.”

In Codex, run /statusline and “Use the picker to toggle and reorder items.” Codex’s command docs say “Available status-line items include model, model+reasoning, context stats, rate limits, git branch, token counters,” and more, saved to tui.status_line in config.toml.

Price: free with Claude Code or Codex.

Watch for: it lives in the terminal session, so it’s gone when the session is. In Claude Code, rate_limits “appears only for claude.ai Pro, Max, and Ultra subscribers, or behind a Claude apps gateway that sets a spend limit for you,” and “Even then, it appears only after the first API response in the session.”

CodexBar

Best for: a menu bar meter for Codex, Claude and most other coding plans, free.

CodexBar is a “Tiny macOS 14+ menu bar app that keeps AI coding-provider limits visible and shows when each window resets.” Its README promises “Per-provider session, weekly, and monthly windows with countdowns to the next reset.” It reads Codex through “OAuth API or local Codex CLI” and Claude through “OAuth API, browser cookies, or CLI PTY fallback; session and weekly usage where available,” and its site counts 111 providers, from Cursor and Copilot to Grok and Antigravity. With Show all accounts in Overview turned on, its Codex docs say it includes “this provider’s cached account cards,” so each saved account gets its own card. It offers “Optional session quota notifications, upcoming-reset warnings for unused Codex quota,” widgets, and a bundled codexbar command that scans local logs for cost. A Linux desktop app, built on Qt 6, has “usage and spending windows, separate settings, desktop notifications, and an optional tray icon.”

It’s free and MIT-licensed, with more than 22,000 stars on GitHub, and version 0.83.0 came out on October 31.

Price: free.

Watch for: there’s no Windows version. The Mac app needs macOS 14 or later, and the Linux app needs “glibc 2.39+, Qt 6.4+, and a separately installed CodexBar CLI.” Some providers work from your browser’s cookies, which you opt into, so macOS may ask for Keychain access, and “Full Disk Access (optional)” is “only required to read Safari cookies/local storage for web-based providers.”

Usagebar

Best for: Claude Code on a Mac, with a warning before you hit the limit, for a one-time price.

Usagebar “tracks your usage limits and remaining session context, directly from your macOS menu bar”: your five-hour and weekly limits, with timers to the next reset, and “alerts at 50%, 75%, and 90% usage so you can wrap up your work before hitting limits.” It “Uses your existing Claude Code credentials via Keychain,” and its FAQ says “Your tokens never leave your machine except to call Anthropic’s usage API directly.”

Price: “$9 — one-time, lifetime updates.”

Watch for: it’s for Claude Code, on a Mac, and it needs the command-line tool: “Usagebar reads credentials from Claude Code (the CLI). You need Claude Code installed and logged in.” Its site doesn’t mention Codex.

ccusage

Best for: seeing how many tokens your agents use, and what that would cost at API prices, across many agents, free.

ccusage is “a local CLI for understanding coding (agent) CLI token usage and estimated costs,” built from the logs agents keep on your computer. Run npx ccusage@latest for reports by day, week, month or session, and its README lists Claude Code, Codex and 16 more agents, OpenCode, Gemini CLI, Antigravity and Grok Build CLI among them. ccusage blocks follows “Claude Code 5-hour billing windows,” with a live view, and ccusage statusline puts the session’s cost, today’s total and the current block in Claude Code’s status line. It’s MIT-licensed, with more than 19,000 stars on GitHub, and version 20.0.48 came out on November 2.

Price: free.

Watch for: it doesn’t read your plan’s limits. Its blocks report measures against a token limit you set, or “Use highest previous block as limit,” and its docs list its limits plainly: “Only analyzes data from your current machine,” tool calls such as web search aren’t counted, and “Costs are estimates and may not reflect actual billing.” On a subscription, the dollar figure answers its own question, “What would I be paying on a pay-per-use plan?”

Orca

Best for: every account’s limits beside your agents, in a free, open-source app.

Orca, from Stably AI, runs coding agents side by side, each in its own worktree, and its status bar shows usage for “Claude Code, Codex, Gemini, OpenCode Go, Kimi Code, MiniMax, Grok, Antigravity, Cursor, and GLM Coding Plan.” Its usage docs list “Current usage against the active account’s plan,” the time to reset for each window, and “A warning chip when you cross 80% of a limit.” Click it for a Usage popover of every provider, “sorted so the tightest limit is first,” in % used or % remaining. “Other configured accounts are visible in the account switcher with their own usage,” Claude and Codex accounts swap in one click, and the phone companion has “the same switcher/usage readout.” It’s free and open source under MIT, on Mac, Windows and Linux.

Price: free.

Watch for: it’s a workspace for running agents rather than a meter on its own. “The status bar always reflects the active account,” and for Codex, “New Codex launches follow the active account; already-running sessions keep the home they started with until you restart them,” per its Codex docs. It checks Claude “about every 15 minutes while it stays visible and focused,” and when you bring it back to the front after five minutes or more.

Enjoy

Best for: every account’s limits for every lab’s agent, and keeping work going when one runs out.

Enjoy is a desktop app for Mac, Windows and Linux that runs Claude Code, Codex, Grok Build, Antigravity, Muse Code, and local models through OpenCode, on the plans you already have. Open the agent settings in the message box, and the usage bars sit under the agent’s name: 5 hours and Weekly for Claude Code, “plus separate weekly limits when your plan has them,” Codex’s limits with its credits and free resets, and the limits Grok Build, Antigravity and Muse Code report, with when each resets. What it adds:

• Every account side by side. The account menu shows each signed-in account’s bars, “including any model-specific limits and credits,” so you can compare before you choose.
• It moves on when one runs out. A line under the bars names the account that takes over, and when the limit hits, Enjoy switches to it and the agent continues right away, or waits for the reset and carries on. Another lab’s agent can pick up the conversation too.
• Codex’s free resets, in one click. When OpenAI gives your account one, you can use it from the same panel.
• A smaller plan, when it would do. On Claude Max, Enjoy watches four weeks of your usage and tells you when a smaller plan would have fit.

Price: Enjoy is one monthly subscription, and the first seven days are free, with a card to start. It includes your phone and any browser and your other computers, and each teammate you invite needs a subscription of their own. See pricing.

Watch for: the bars are in the computer app only: “The web app shows Usage unavailable, and the iPhone app doesn’t show usage.” Enjoy checks when you open the agent settings or the account menu, and after each turn, and “It doesn’t check usage between turns.” OpenCode shows Usage unavailable, because “OpenCode’s usage depends on the providers you set up in it.” On Windows, saved accounts need Developer Mode turned on, except in projects inside WSL. Enjoy isn’t open source, it’s a monthly subscription on top of your AI plans once the free week ends, and waiting for a reset needs your computer on, with Enjoy open, when the usage resets.

When the built-in commands are enough

For one account with one lab, they are. /usage and Settings > Usage show Anthropic’s own numbers, and /status and Settings → Usage show OpenAI’s, at no extra cost. The trackers above read the same numbers from the same companies, and ccusage estimates from your logs. Check before a long task, and watch the week as well as the five hours: Anthropic’s weekly limit resets on a fixed day and time for your account, and OpenAI’s five-hour limit can stop you while you still have weekly usage left.

A tracker earns its place when you check often enough that opening a page is a chore, when you pay for both labs, or when you have more than one account. If what you need is more usage rather than a better view of it, our guides to Claude Code’s limit and Codex’s cover resets, credits and the other ways through.

How we compared

We read Anthropic’s Claude Code documentation on costs, commands, the status line and VS Code, its help center articles on usage limits and its pricing; OpenAI’s Codex pricing and command docs and its help center articles on usage; the READMEs, sites and docs of CodexBar, ccusage and Orca; and Usagebar’s site, on November 2, 2026, and every one is linked below. Licenses, star counts and release dates come from GitHub that day. We quote each maker’s own words for its limits and leave out what it hasn’t documented, and we didn’t test the third-party tools ourselves. These tools change weekly, so we recheck this guide against its sources at least once a month, and the date at the foot of the page says when we last did.