Claude Code Approaching 5 Hour Limit: What It Means

If Claude Code says you're approaching your 5 hour limit, it means you've used most of your plan's current session window, not that something broke. On a Pro or Max subscription, Claude Code draws from a rolling five-hour allowance and a weekly allowance at the same time, and both are shared with your chats on claude.ai.

Here's the short version, based on Anthropic's Claude Code docs and Help Center as of October 2026:

• Two windows apply at once. A five-hour session limit and a weekly limit. Max plans also have a separate weekly limit for Fable.
• Claude and Claude Code share the pool. Long chats on claude.ai eat into your coding time, and the reverse.
• The warning comes first. You'll see something like You've used 85% of your session limit · resets 3:45pm before the window runs out.
• Check where you stand with /usage inside Claude Code.
• When you run out, wait for the reset, let Claude Code continue automatically, turn on usage credits, or sign in with a Console API key.
• Stretching usage comes down to model choice, /clear between tasks, /compact on long ones, and smaller subagents.

The rest of this guide walks through each one.

How Claude Code usage limits work

Anthropic doesn't sell Claude Code by the message. Subscription plans include a rolling usage allowance, and every request draws from it. How fast you drain it depends on the model, how much context each turn carries, and how much work the agent does per turn.

The five-hour session limit and the weekly limit

The Pro and Max help article lists the limit types plainly: a five-hour session limit and a weekly limit, plus a separate weekly limit for Fable on Max.

The detail most people miss is in the error reference: usage counts against the session and weekly allowances at the same time. A five-hour reset gives you a fresh session window, but whatever you used still sits on your weekly total. Anthropic also warns that one burst of heavy activity, such as a large workflow fanout, can exhaust the weekly allowance before the session window even resets.

Anthropic doesn't publish a fixed number of prompts or tokens per window, and we don't guess at one here. The live numbers are in your account.

The limit is shared with claude.ai

Per the same help article, Claude and Claude Code draw from the same limits, so activity in either tool counts toward both. If you spent the morning in a long research chat, your afternoon coding session starts with less room.

On Team and Enterprise plans, the costs guide says each member's Claude Code usage draws from a per-seat allowance that resets on a rolling five-hour window and a weekly window. That allowance is shared with Claude chat and Cowork, and its size depends on whether the seat is Standard or Premium.

Which limit you hit decides what helps

Claude Code shows four different messages when a window runs out:

The session and weekly limits are shared across all models, so switching models doesn't restore access. The Opus and Sonnet limits are per family. One catch from the docs: each model keeps its own prompt cache, so the first request after a switch re-reads the whole conversation with no cache hits. Our guide to changing the model in Claude Code covers the picker and the settings that stick.

What "approaching 5 hour limit" actually means

The warning is an early heads-up, not a block. The error reference says Claude Code can warn you before a window runs out, with a message such as You've used 85% of your session limit · resets 3:45pm. The help article puts it more simply: you'll see warnings as you approach any limit.

When you see it, three things are worth doing:

1. Note the reset time. It's printed right in the warning.
2. Finish what's in flight. Wrap up the current change and commit, so a cutoff doesn't land mid-refactor.
3. Drop to a lighter model for whatever is left, if the remaining work is routine.

Don't confuse it with the context warning

Claude Code has a second warning that looks similar and means something else. A context or auto-compact warning tells you the conversation has grown close to the model's context window, where Claude Code summarizes older history to free space. The costs guide is explicit that this is not a usage limit. It's about how much one conversation can hold, the same distinction we draw in our Claude upload limit guide between file caps and the context window.

What happens if the limit hits mid-response

The help article describes a small wrap-up allowance. If you reach the five-hour limit mid-response, Claude Code may briefly keep working to finish the current step, and you'll see Usage limit reached · wrapping up.

The fine print matters here:

• The allowance is small and capped, and it counts toward your weekly limit.
• It covers a response already in progress, not new messages.
• On Pro it applies up to once per weekly period. On Max it can apply each time you reach a five-hour limit, within the weekly limit.
• Anthropic says it's rolling out gradually and may not be on your account yet.

How to check your Claude Code usage

Run /usage. The commands reference describes it as showing session cost, plan usage limits, and activity stats. /cost and /stats are aliases for the same screen.

On a subscription, the useful parts are the plan usage bars and, on Pro, Max, Team, or Enterprise, a breakdown of what counts against your plan limits:

• Attribution: recent usage split across skills, subagents, plugins, and individual MCP servers, each as a percentage.
• Recommendations: one row per suggested change, such as trimming a large CLAUDE.md, shown for behaviors behind 10% or more of your recent usage.

Press d or w to switch between the last 24 hours and the last 7 days. One caveat from the docs: the breakdown is computed from local session history on that machine, so it doesn't include claude.ai or your other devices. The plan bars are the authoritative view.

Two other places show the same allowance:

• A custom status line. Add the rate_limits fields to watch your remaining allowance continuously.
• claude.ai Settings > Usage. The usage limit best practices article says this page shows how much of the five-hour session limit you've used, the time remaining, and when the weekly limit resets.

What to do when you hit the limit

You have more options than waiting, though waiting is often the right call.

Let Claude Code continue after the reset

In an interactive session signed in with a claude.ai subscription, Claude Code waits in the open session and continues the task on its own after the limit resets. The interactive mode docs say this is on by default from v2.1.234. You'll see two lines at the bottom of the session:

Usage limit reached · limit resets 3:45pm
Continuing automatically at 3:45pm · Esc to cancel

Keep the session open. At the reset, Claude Code prompts Claude to pick up where it stopped rather than resending your last message. It still asks for permissions as usual, so a task can stall on a prompt while you're away.

It won't start the wait on its own when the reset is more than 24 hours away, which is common with a weekly limit. In that case, run /rate-limit-options and pick Wait here, then continue automatically. To turn the behavior off, switch off Continue automatically at usage limit in /config.

Turn on usage credits

Usage credits let you keep working past your plan's limit on pay-as-you-go billing. Anthropic's usage credits article says they apply to both Claude conversations and Claude Code terminal usage, and lets you set a monthly spending cap.

Inside Claude Code, run /usage-credits:

On Pro and standard Team or Enterprise seats, Fable runs on usage credits from the start, according to the Fable plan article. On Max and premium seats, Fable can use up to half of your weekly limit before you need credits.

Fall back to an API key

You can also switch to a Claude Console account and pay standard API rates. Two details from the help article are worth knowing before you do:

• If ANTHROPIC_API_KEY is set in your environment, Claude Code uses that key instead of your subscription, and you're charged API rates. If your limits seem to vanish or your bill looks odd, check for a stray key in your shell config.
• The monthly API credits included with some Max plans can't be used in Claude Code.

With API-key billing there's no session window to wait out. Usage is metered per request.

To go back to your subscription only, logging out isn't enough on its own, because the environment variable survives a re-login. Anthropic's API key help article gives the steps:

1. Run unset ANTHROPIC_API_KEY in your shell. If the key is exported in a profile such as ~/.zshrc, delete that line too, then reload the profile or open a new terminal.
2. Sign back in with your Pro or Max account: claude auth logout then claude auth login from the terminal, or /logout and /login inside a session.
3. Run /status in Claude Code to confirm which authentication method is active.

Also decline API credit offers if Claude Code shows them, so you don't drift back onto API billing.

Upgrade the plan

If you hit the limit every week, the help article's own advice is to move from Pro to Max 5x, or from Max 5x to Max 20x. Compare tiers on claude.com/pricing. Our Claude Max vs Pro breakdown explains what the 5x and 20x multipliers apply to.

How to make your Claude Code usage last longer

Every tip here comes from Anthropic's own costs guide. The common thread: tokens scale with context, so less context per turn means more turns per window.

Match the model to the job. The docs say Sonnet handles most coding tasks well and costs less than Opus, and Anthropic's models and limits article says Opus "uses meaningfully more of your quota." A common pattern is to plan with Opus and execute with Sonnet, which is what the opusplan setting does.

Use /clear between unrelated tasks. Stale context costs tokens on every later message. Run /rename first if you want to come back to the old session with /resume.

Use /compact on long tasks. It summarizes the conversation to free space. You can steer it: /compact Focus on code samples and API usage tells Claude what to keep.

Keep CLAUDE.md short. It loads into every session. Anthropic suggests staying under about 200 lines and moving workflow-specific instructions into skills, which load only when used.

Send noisy work to subagents, on a smaller model. Test runs, log processing, and doc fetching can flood your main context. A subagent keeps that output in its own context and returns a summary. Its requests still draw on your usage, so set model: haiku for simple subagent jobs. Be careful with agent teams: the docs say they use roughly 7x more tokens than a standard session when teammates run in plan mode.

Trim what loads every turn. Disable unused MCP servers with /mcp, lower effort with /effort on simple tasks, and run /context to see what's taking space.

Coming from OpenAI's tool? Our Codex usage limits guide covers how its five-hour and weekly windows compare.

Where Taku fits

A lot of Claude Code usage goes on setup rather than the task: getting a repo running, wiring tools, re-explaining a project every session. If that's where your window goes, Taku mirrors a working AI setup into a desktop workspace and runs it, so you start from a configuration someone already got working. Taku is in Beta, and the Mac app is available now.

FAQ

What does "approaching 5 hour limit" mean in Claude Code?

You've used most of your current five-hour session window. Claude Code shows a warning such as You've used 85% of your session limit with the reset time, before it blocks requests.

What are the Claude Code usage limits on Pro and Max?

Both plans have a five-hour session limit and a weekly limit, shared with claude.ai. Max also has a separate weekly limit for Fable. Anthropic doesn't publish fixed message counts; check /usage or claude.ai Settings > Usage for your live numbers.

Does switching models fix a Claude Code usage limit?

Only for the model-specific ones. After "You've hit your Opus limit" or "Sonnet limit", switching to another family keeps you working. The session and weekly limits cover all models, so /model won't help.

Is there a separate Claude Code max Opus limit?

Claude Code can show a "You've hit your Opus limit" message that applies only to Opus-family requests. Anthropic doesn't publish its size. Switching to Sonnet with /model keeps you working until it resets.

When does the 5 hour limit in Claude reset?

The warning and the limit message both print the reset time. /usage and claude.ai Settings > Usage show it too, along with when your weekly limit resets.

Can I keep using Claude Code after hitting the limit?

Yes. Let Claude Code continue automatically after the reset, turn on usage credits with /usage-credits, or sign in with a Console API key and pay API rates.

Key points

• Claude Code on Pro and Max runs on a five-hour session limit and a weekly limit, counted at the same time.
• The allowance is shared with claude.ai, and on Team and Enterprise with chat and Cowork too.
• The "approaching" warning shows your percentage used and reset time; it's different from the context warning.
• /usage shows plan bars and what's using your allowance.
• At the limit: wait and auto-continue, use /usage-credits, switch to an API key, or upgrade.
• Sonnet for routine work, /clear, /compact, a short CLAUDE.md, and Haiku subagents stretch a window furthest.