What is the Claude 5-hour session limit?

The five-hour session limit is a rolling window that opens with your first message and caps how much Claude you can use inside it. It is not a daily allowance, it is not the weekly cap, and it does not start when you log in. When you hit it, access pauses until that window rolls off, which happens five hours after the message that opened it.

What the five-hour limit actually is

Claude meters subscription usage on two clocks at once. The first is the five-hour session window: a rolling block that starts when you send your first message and governs how much you can do in a short burst. The second is a set of weekly caps layered on top. The five-hour window is the one most people hit first, because it is the one that bites during a single heavy afternoon.

The word "session" is misleading if you read it the way apps usually use it. It is not a login, not a conversation, and not a Claude Code process. It is a five-hour timer that started the moment you first messaged Claude, and it keeps running whether you are active or idle. Closing the tab, quitting Claude Code, or opening a brand new chat does not end it and does not start a fresh one.

What starts the window, and the batching trap

The window starts with your first message. Not a login, not opening the app, and not a fixed slot on a shared schedule. That is the single most useful fact about it, because it means the clock is anchored to when you start rather than to everyone else's day.

One wrinkle on the exact minute: Anthropic's usage documentation notes that reset times land on the top of an hour, with the clock counted from the hour in which you sent that first message, so a window you open at 8:55am can end at 1:00pm rather than 1:55pm. Read the exact time off your usage screen rather than counting minutes from your first message.

The window is also rolling rather than fixed. When one expires, the next begins with the next message you send, so your windows drift through the day depending on when you actually work. That has direct practical consequences:

• A throwaway question opens a real window. Ask Claude something trivial at 8:55am and that is the same window governing your 9:30am heavy refactor, and it expires earlier than one you would have opened at 9:30.
• Batch heavy work into one block. Opening the window when you are actually ready gives you the most of it for work that matters.
• Avoid scattering small asks across a morning. Three stray questions between 8:00 and 11:00 can straddle two windows and leave you starting the real task on a half-spent one.

None of this is about the time of day. Anthropic removed the peak-hours limit reduction on Claude Code for Pro and Max accounts on May 6, 2026, so as of August 2026 there is no busy-hours penalty to schedule around. See the best time to use Claude.

What "session limit reached" means

When Claude tells you the session limit is reached, it means the current five-hour window is spent. It does not mean your account is out of usage for the day, the week, or the month, and it is not a penalty or a strike against your account. Access pauses and resumes automatically when the window rolls off, with nothing for you to do.

Options while you wait:

• Wait it out. The refill is automatic and the usage screen shows the exact time. See how to check your Claude usage.
• Work out which clock you hit. A session limit and a weekly limit look similar in the moment but mean very different waits: hours versus days.
• Upgrade, or buy usage credits. Paid plans can purchase additional usage rather than wait. Whether that is worth it depends on how often you land here; Claude Max vs Pro covers the tier question.
• Spend the pause on the half of the work Claude was not doing. Review the last diff, write the test, read the API docs, decide what you actually want next. A prompt drafted during the wait is worth more than the one you would have improvised.

When the session limit resets

The window resets five hours after it opened, which is five hours after your first message of that window and not five hours after you hit the limit. If you opened it at 9:00am and ran out at 10:30am, you are waiting until 2:00pm, not until 3:30pm. That asymmetry surprises people, and it usually works in your favor.

Session limits reset every five hours, continuously, with no action needed and nothing to reset by hand. There is no midnight reset involved anywhere in this: see the Claude daily usage limit question for why the daily mental model does not fit. To see your own reset time, open Settings > Usage in the web or desktop app, or run /usage in Claude Code, which returns immediately without interrupting a response and shows a progress bar and next reset per window. The time shown is in your local timezone, which is the honest answer to the question covered in when Claude limits reset.

How it interacts with the weekly caps

The two clocks run independently, so you can hit one with plenty of room on the other. One intense afternoon can exhaust a five-hour window while your weekly caps are barely touched. Weeks of steady moderate use can walk you into a weekly cap without your ever seeing a session limit.

On the weekly side there are actually two ceilings: one across all models, and a separate one for Opus. Hitting the Opus cap while the general cap still has room is common if you default to the most capable model for routine work. Full mechanics are in Claude weekly usage limits, explained, and the early-warning banner is covered in approaching weekly limit.

Two more things worth knowing. Every Claude surface draws from the same pool: claude.ai, Claude Desktop, and Claude Code all count against the same windows and caps, so a heavy terminal day and a heavy browser day compound. And on May 6, 2026, Anthropic doubled Claude Code's five-hour rate limits for Pro, Max, Team, and seat-based Enterprise plans, at the same time as removing the peak-hours reduction. If your sense of how far one window goes was formed before that, it is out of date.

The windows you never fill are the interesting ones

This follows directly from how the window works: it is a ceiling, not a balance. Whatever you do not use before a window rolls off is gone, and the next one refills to the same level whether you used all of it or none of it. Most people leave most of their five-hour windows almost entirely untouched.

Tokens for Good captures that recurring surplus. Your Claude Code claims a queued nonprofit, researches its real-world impact against a fixed methodology with citations, and submits a structured report. Every organization is researched twice by independent contributors, then validated, consolidated, scored deterministically, and human-reviewed before it reaches the public directory. It runs on the subscription you already pay for with no separate API cost, and it can run on a schedule in the background, which means it quietly fills windows you were never going to use. To be clear, "tokens" here means AI model tokens, not crypto: no coin, no wallet, no blockchain. See how the research works or the docs, which take about a minute.

Frequently asked questions