# How to check your Claude Code usage

Authors: AutoRouter
Published: 2026-10-29T14:00:00.000Z
Reading time: 8 min
Canonical: https://autorouter.com/blog/check-claude-code-usage
Part of: How to cut Claude Code and Codex costs (https://autorouter.com/blog/cut-claude-code-and-codex-costs.md)
Publisher: AutoRouter, the Mac app for Codex and Claude Code (https://autorouter.com)

> Run /usage in Claude Code to see your plan usage bars, a breakdown of what used your plan, and, for API users, a session token and cost estimate. Add a status line to keep plan and context usage on screen. AutoRouter, the Mac app for Codex and Claude Code, also shows the model and cost of every turn.

To check your Claude Code usage, run `/usage` inside a session. On a Pro, Max, Team or Enterprise plan it shows your plan usage bars and a breakdown of what used them, and for API users its Session block shows tokens and an estimated cost, according to Anthropic's [cost docs](https://code.claude.com/docs/en/costs) (checked October 26, 2026). For a number that stays on screen, add a status line, and for a picture of the model and cost behind every turn, [AutoRouter, the Mac app for Codex and Claude Code](https://autorouter.com), shows both on each turn.

AutoRouter wrote this post. Everything about Claude Code comes from Anthropic's documentation, linked where it is used.

## How do you check Claude Code usage?

Type `/usage` at the Claude Code prompt. Anthropic's [cost guide](https://code.claude.com/docs/en/costs) names it as the place to check your current token usage, and the screen has three parts:

- **Session block**: total cost, API and wall time, lines changed, and a `Usage by model` line with input, output and cache tokens for each model in the session.
- **Plan usage bars**: how much of your plan's limits you have used, on Pro, Max, Team and Enterprise.
- **Usage breakdown**: which skills, subagents, plugins and MCP servers used your plan, with suggestions for using less.

The Session block resets when `/clear` starts a new session, so the next session's total starts at $0. On Claude Code v2.1.251 or later it also gains a `Prompt cache (main)` line showing how much of your input came from cache and how many requests missed it.

## Is the cost in /usage what you actually pay?

No, unless you pay per token at list price. Claude Code computes the dollar figure locally from token counts at list price, and the docs call it an estimate, pointing API users to the [Claude Console usage page](https://platform.claude.com/usage) for billing. For Pro and Max subscribers, usage is included in the plan, so the session cost "isn't relevant for billing purposes" ([docs](https://code.claude.com/docs/en/costs)).

What counts for subscribers is the plan usage bars. Organizations on contracted rates can set the `modelPricing` managed setting so the figure matches their contract; the `Total cost` line then notes that it uses your organization's configured rates.

## What does the usage breakdown show?

The breakdown shows what used your plan over the last 24 hours or 7 days; press `d` or `w` to switch. It covers:

| Row             | What it tells you                                                            | Version needed      |
| --------------- | ---------------------------------------------------------------------------- | ------------------- |
| Attribution     | Share of recent usage from each skill, subagent, plugin and MCP server       | Any current version |
| Recommendations | A suggested change, shown when it applies to at least 10% of your usage      | Any current version |
| Loops           | The heaviest `/loop` and other scheduled tasks, with runs and tokens per run | v2.1.242 or later   |
| `f to fix`      | Has Claude Code apply the recommendations it can                             | v2.1.305 or later   |

A recommendation's share is how much of your usage the row applies to, not how much the change saves, and one request can count toward several rows. The figures are approximate and come from session history on this machine, so usage from your other computers or from claude.ai isn't included.

If `/usage` can't reach the usage service, it shows the last bars it loaded in the past 60 minutes with a `Showing last-known usage` note; press `r` to retry. A message such as `rate limited until 3:14am` means Claude Code can't fetch your limits again until then, not that your work is limited.

## How do you keep usage on screen while you work?

Add a status line, a script whose output Claude Code shows at the bottom of the screen. Run `/statusline` with a description, such as `/statusline show context and plan usage`, and Claude Code writes the script for you. The status line runs locally and uses no API tokens, according to the [status line docs](https://code.claude.com/docs/en/statusline).

Claude Code passes the script JSON about the session. The fields that matter for usage are `cost.total_cost_usd`, `context_window.used_percentage`, and `rate_limits.five_hour.used_percentage` and `rate_limits.seven_day.used_percentage`, which appear for claude.ai Pro and Max subscribers after the session's first response. A small script that shows all three:

```bash
#!/bin/bash
# ~/.claude/statusline.sh
jq -r '"[\(.model.display_name)] context \(.context_window.used_percentage // 0)% | 5h \(.rate_limits.five_hour.used_percentage // 0)% | 7d \(.rate_limits.seven_day.used_percentage // 0)%"'
```

The script uses [`jq`](https://jqlang.org/), which you may need to install. Make it executable with `chmod +x ~/.claude/statusline.sh`, then point Claude Code at it in `~/.claude/settings.json`:

```json
{
  "statusLine": {
    "type": "command",
    "command": "~/.claude/statusline.sh"
  }
}
```

While a turn runs, the spinner line also shows a live count of the output tokens that turn has produced, as in `(10m 27s · ↓ 5.5k tokens)`. It resets at the start of each turn, so it won't match the session totals in `/usage`. To see which model ran each finished turn and what it cost, export Claude Code's telemetry, as how to see which model handled each turn shows.

## Which tool should you use to track Claude Code usage?

It depends on who pays and what you want to know. Anthropic's tools cover different needs:

| You want to see                         | Use                                                                      | Who it's for                        |
| --------------------------------------- | ------------------------------------------------------------------------ | ----------------------------------- |
| How much of your plan is left           | `/usage` plan bars, or the `rate_limits` status line fields              | Pro, Max, Team, Enterprise          |
| What used your plan                     | `/usage` breakdown                                                       | Pro, Max, Team, Enterprise          |
| Tokens and estimated cost this session  | `/usage` Session block or `cost.total_cost_usd`                          | API users                           |
| How you work, and what to change        | `/insights`                                                              | Any plan or provider                |
| Usage credits spent this month          | `/usage` usage-credits row, `/usage-credits`                             | Plans with usage credits on         |
| Spend per developer                     | Org analytics spend report, Console dashboard, Claude Code Analytics API | Team, Enterprise, API organizations |
| Live per-user metrics in your own tools | OpenTelemetry export                                                     | Every setup                         |

`/insights` analyzes up to 200 sessions on this machine that it hasn't seen before and writes an HTML report to `~/.claude/usage-data/report.html`. The analysis runs on your account, so its tokens count against your plan or API usage.

## What do the usage limit messages mean?

Each message names a different limit. "You've hit your session limit" and "You've hit your weekly limit" are plan windows shared across all models, so switching models with `/model` won't get you going again. After "You've hit your Opus limit" or "You've hit your Sonnet limit", switching to a model outside that family does keep you working ([docs](https://code.claude.com/docs/en/costs)).

On Claude Code v2.1.234 or later you can wait and have the interrupted task continue after the reset, and with usage credits on, `/usage-credits` lets you keep going past the limit. [How Claude Code usage limits work](/blog/claude-code-usage-limits) explains the five-hour and weekly windows in detail.

## How do you make your usage go further?

Keep context small and match the model to the job. Anthropic's [cost guide](https://code.claude.com/docs/en/costs) lists these habits:

- Run `/clear` between unrelated tasks, and `/compact` with instructions about what to keep.
- Use Sonnet for most coding and save Opus for complex architecture or multi-step reasoning; give simple subagents `model: haiku`.
- Lower `/effort` for simple tasks.
- Run `/context` to see what fills the context window, and `/mcp` to turn off servers you aren't using.
- Move workflow instructions from CLAUDE.md into skills, which load only when used.

For scale, Anthropic reports enterprise deployments averaging about $13 per developer per active day and $150 to $250 per developer per month, with 90% of users under $30 per active day.

## How does AutoRouter show usage for each turn?

AutoRouter shows the model, the cost and the reason on every turn while it is the model selected in Claude Code's `/model` picker, so you can see where your plan went turn by turn. It works out how hard each turn is and only considers models that clear the quality bar for it. Hard work goes to the most capable model your plans cover, and everything else goes to the model that costs you least, so your Claude plan goes to the work that needs it and lasts longer.

When your Claude plan hits its limit, AutoRouter Pro keeps those turns going from its $20 of included usage each month until the plan resets, and asks you to top up if that runs out too. AutoRouter is in beta, runs on macOS 14 or later, and needs Node 22.18 or later for Claude Code. [How AutoRouter picks a model for each turn](/blog/how-autorouter-picks-a-model) covers the routing in full, and [every way to choose a model in Claude Code](/blog/claude-code-model-settings) covers picking one yourself.

## Questions

### Does /usage show usage from claude.ai or my other computers?

No. The plan usage breakdown is computed from session history on the machine you run it on, so claude.ai and other devices aren't included. The plan usage bars come from your account's limits.

### Why does /usage show a cost if I'm on a subscription?

The Session block is meant for API users and computes an estimate at list price. On Pro and Max, usage is part of the plan, so watch the plan usage bars instead.

### Does the status line use up my plan?

No. The status line script runs locally and uses no API tokens. `/insights` is different: its analysis counts against your plan or API usage.

### When does the usage in /usage reset?

The Session block resets when `/clear` starts a new session. Plan limits reset on their own windows; the `rate_limits` status line fields include a `resets_at` time for each window.

## Sources

- [Manage costs effectively](https://code.claude.com/docs/en/costs), checked October 26, 2026
- [Customize your status line](https://code.claude.com/docs/en/statusline), checked October 26, 2026
- [Claude Console usage page](https://platform.claude.com/usage), checked October 26, 2026

## Read next

- [What common Claude Code tasks cost, measured on each model](https://autorouter.com/blog/claude-code-task-costs-measured.md): AutoRouter ran five everyday Claude Code tasks three times each on Claude Opus 5.5, Sonnet 5.5 and Haiku 5.5 in October 2026. At Anthropic's list prices, an average run cost $0.110 on Opus 5.5, $0.049 on Sonnet 5.5 and $0.0044 on Haiku 5.5, and all 45 runs passed their checks. Writing the prompt cache was the largest part of every model's cost.
- [How to cut Claude Code and Codex costs](https://autorouter.com/blog/cut-claude-code-and-codex-costs.md): Claude Code and Codex costs come down to three things: which model runs each turn, how much context each request carries, and whether you pay through a plan, credits or an API key. Run routine turns on a smaller model, keep sessions short, and spend the plan you already pay for before paying per token. AutoRouter, the Mac app for Codex and Claude Code, makes the model choice on every turn and shows each turn's model, cost and reason.
- [How to use Claude Code and Codex without a subscription](https://autorouter.com/blog/no-subscription-coding-agents.md): Claude Code runs on a Claude Console account and Codex runs on an OpenAI API key, and both bill per token with no ChatGPT or Claude plan. AutoRouter, the Mac app for Codex and Claude Code, is a third option: AutoRouter Pro costs $20 a month with $20 of included usage and bills paid models at the provider's price plus 10%. On the same models, a direct API key costs less, so Pro comes out ahead only when cheaper models take enough of your turns.

Tags: [Claude Code](https://autorouter.com/blog/tags/claude-code.md), [Costs](https://autorouter.com/blog/tags/costs.md), [Guides](https://autorouter.com/blog/tags/guides.md)

Drafted with AI from the sources linked above, then checked and edited by the AutoRouter team.