Claude Code rate limits: usage limits per plan, checking usage, and what to do when you hit one

Claude Code rate limits on a Claude subscription are usage limits: a session limit that resets every five hours and a weekly limit, shared with everything else you do in Claude. Run /usage to see how much of each you have used and when they reset. When you hit one, Claude Code blocks new requests until the reset, unless you turn on usage credits, upgrade, or switch to pay-as-you-go API billing.

This guide follows Anthropic's help center and the Claude Code docs as read on 26 October 2026. Anthropic describes the allowances as multiples between plans rather than fixed message counts, so this page does too.

What are the Claude Code usage limits on each plan?

According to Anthropic's Pro and Max help article: "Both Pro and Max plans have a five-hour session limit and a weekly limit. Max plans also have a separate weekly limit for Fable. These limits are shared across Claude and Claude Code, meaning all activity in both tools counts against the same limits."

Here is how the plans compare, with prices from claude.com/pricing as of October 2026:

A few details matter for planning:

• Weekly resets are fixed. The Max plan article says the weekly limit resets at a fixed time each week assigned to your account, regardless of when you started using Claude. You can see it in Settings > Usage.
• Usage counts against both windows at once. The Claude Code error reference warns that one heavy burst, such as a large workflow fanout, can use up the weekly allowance before the five-hour window resets.
• What you do changes how fast it drains. The pricing page says it depends on the length and complexity of your conversations, the model you choose and the features you use, so there is no fixed message count.
• Anthropic can add other caps. The Max article says it may limit usage "in other ways, such as weekly and monthly caps or model and feature usage, at our discretion."

For costs beyond the limits and the full plan breakdown, see our Claude Code pricing guide.

How do API keys and cloud providers handle limits?

If you sign in with a Claude Console API key, or run through Amazon Bedrock, Google Cloud's Agent Platform or Microsoft Foundry, there is no five-hour or weekly window. You pay per token, and Anthropic's models, usage and limits article puts it plainly: "No hard stop; the account is charged for what it uses."

You can still be rate limited. API organizations have per-minute limits, and Claude Code traffic counts toward them. When you hit one, you see:

API Error: Request rejected (429) · this may be a temporary capacity issue. If it persists, check https://status.claude.com.

For teams, the costs page suggests per-user tokens-per-minute (TPM) and requests-per-minute (RPM) allocations that shrink as the team grows, because fewer people run Claude Code at the same moment:

Watch for one trap. Per the help center, if ANTHROPIC_API_KEY is set in your environment, Claude Code uses that key instead of your subscription, which means API charges and API rate limits. Our environment variables guide explains how to check and unset it.

How do you check your Claude Code usage?

Run /usage. Per the commands reference, it shows session cost, plan usage limits and activity stats. /cost and /stats are aliases. On Pro, Max, Team and Enterprise plans it also shows a breakdown of what used your limits: skills, subagents, plugins and MCP servers as percentages, plus suggestions such as trimming a large CLAUDE.md. Press d or w to switch between the last 24 hours and 7 days. The breakdown covers only this machine, not claude.ai or other devices.

/usage or /status? One help-center article still says to monitor your allocation with /status. In current Claude Code, /status shows version, model, account and connectivity, while plan limits and reset times are in /usage. /status is still the place to confirm which credential is active.

Watch the warnings. Before a window runs out, Claude Code shows messages such as You've used 85% of your session limit · resets 3:45pm.

Add it to your status line. For Pro and Max subscribers, the status line receives a rate_limits object with five_hour and seven_day windows, each with used_percentage and resets_at. Anthropic's status line docs show a script that prints both percentages next to the model name, and our status line guide covers setup.

Check on the web. Settings > Usage on claude.ai shows where you stand and your next reset time. In the Claude Desktop app, click the usage ring next to the model picker.

One confusing message: rate limited until 3:14am inside /usage means Claude Code can't refresh your usage numbers until then. It does not mean your work is limited.

What happens when you hit a limit?

You see one of these, each with its reset time:

You've hit your session limit · resets 3:45pm
You've hit your weekly limit · resets Mon 12:00am
You've hit your Opus limit · resets 3:45pm
You've hit your Sonnet limit · resets 3:45pm

The session and weekly limits apply to all models, so switching models won't help. The Opus and Sonnet limits apply only to that family, so /model to a different family keeps you going.

Two features soften the stop:

• Wrap-up allowance. If you hit the five-hour limit partway through a response, Claude Code may keep working briefly to finish the current step, showing Usage limit reached · wrapping up. The wrap-up allowance article says it needs Claude Code 2.1.277 or later, is small and capped, counts toward your weekly limit, applies up to once per weekly usage period on Pro, and is rolling out gradually.
• Automatic continue. From v2.1.234, in an interactive session signed in with a subscription, Claude Code waits in the open session and picks the task back up after the reset. It is on by default. Keep the terminal open, press Esc to cancel, or turn off Continue automatically at usage limit in /config. It doesn't apply to -p runs, background sessions or API keys.

Not every error is your quota. API Error: Server is temporarily limiting requests (not your usage limit) is a short server-side throttle that Claude Code retries, and a 529 Overloaded error "is not your usage limit and doesn't count against your quota." For both, wait and check status.claude.com.

What can you do when you keep hitting limits?

Anthropic's options, from the help center and the interactive mode docs:

1. Wait for the reset. The message tells you when.
2. Run /rate-limit-options. It offers to wait and continue automatically, add usage credits, or upgrade.
3. Turn on usage credits. Run /usage-credits, or go to Settings > Usage. Requests then continue at standard API rates, up to any monthly spend limit you set. Team and Enterprise members without billing access send the request to their admin instead. Monthly API credits included with Max plans can't be used in Claude Code.
4. Upgrade. Pro to Max 5x, or Max 5x to Max 20x.
5. Switch to a Console account and pay per token for heavy sprints.

How do you make your usage last longer?

Every turn resends the whole conversation, so the biggest savings come from keeping context small:

• Clear between tasks. /clear starts fresh; /compact summarizes when you need to keep going. Neither resets your limit, but both make later turns cheaper.
• Match the model to the job. Anthropic's support article notes that Opus "uses meaningfully more of your quota." Use Sonnet for routine work, or opusplan to plan with Opus and execute with Sonnet. See our models guide.
• Lower the effort level for routine tasks.
• Point at files instead of pasting them, and trim logs to the relevant lines.
• Keep CLAUDE.md lean, since it loads on every turn.
• Plan before big changes. A plan costs far less than a wrong diff you regenerate.

Our context window guide and cost guide go deeper.

Prompt reuse helps here too: a precise, tested prompt wastes fewer turns than one retyped from memory. Kipt keeps your prompts and Agent Skills as Markdown in a private GitHub repo you own, with {{name}} blanks you fill before copying.