Claude Usage Limits Explained: Pro, Max and Weekly Limits

You are halfway through a task and Claude stops. 5-hour limit reached in chat, or You've hit your weekly limit in Claude Code, with a time attached. That is the Claude usage limit, and it comes in layers.

A session limit resets every five hours. Paid plans add a weekly limit on top, and Max adds a separate weekly limit for Fable. Everything you do in Claude draws on them: chat on claude.ai, a task in Claude Desktop, a session in Claude Code.

Anthropic says there is no fixed message count. It gives multiples instead: Pro is at least 5x Free per session, Max 5x is 5x Pro, Max 20x is 20x Pro. So the useful questions are which limit you hit, when it comes back, and what to do until then.

This guide answers them from Anthropic’s own pages, linked as we go and checked October 26, 2026. Anthropic changes these limits, so treat every number as a snapshot. We make Context Board, a live board that can keep both limits on screen, and it gets one short section near the end.

Claude usage limits by plan

Anthropic’s pricing page says how much you can do “depends on the length and complexity of your conversations, the model you choose, and the features you use, so there’s no fixed message count.” What it does state is how plans compare with each other. This is the picture as of October 26, 2026.

The pages we read state the multiples per five-hour session and give no weekly multiple. A few notes on the rows:

• Free: one limit, the five-hour session. Claude tells you when you’ve reached it.
• Claude Pro limits: the session limit plus a weekly limit across all models. How many messages that is varies with message length, attached files, the length of your conversation and the model or feature you use. Fable isn’t in the plan’s limits, more on that below.
• Claude Max limits: the same two limits as Pro, with five or twenty times Pro’s per-session allowance, plus the separate Fable limit.
• Team: Standard seats include 1.25x the Pro plan’s per-session allowance and Premium seats 6.25x, each with a weekly limit across all models. Limits are per member, so one person hitting theirs doesn’t affect the others, per the Team plan article.
• Enterprise: the current plan is usage-based. The seat fee covers access, there are no plan or seat-level usage limits, and usage is billed at API rates, per the Enterprise plan article. Older seat-based Enterprise plans keep Standard and Premium seats, and those do have limits.

Anthropic also says plans and prices are subject to change at its discretion, and that it may limit usage in other ways, such as weekly and monthly caps or model and feature usage. That is why this page says when it was checked.

The five-hour session limit

The session limit is the short one. It is an allowance for a rolling five-hour window, and Free, Pro, Max and Team plans all have one. Claude’s message calls it the 5-hour limit. Use it up and you wait for the window to roll over, keep going on usage credits if your plan allows, or move up a plan.

On a paid plan, picture two meters running at once. Each message counts against the five-hour meter and the weekly meter together, and you stop when either one runs out. Claude Code’s docs put it the same way. Usage counts against the session and weekly allowances at the same time.

Anthropic’s usage limit best practices list the factors that use up your limits.

• Message length and file attachment size
• The length of the current conversation
• Tools such as Research and web search
• Model choice and effort level
• Artifact creation and use
• Multi-step tasks, like running code, creating files or browsing websites

Anthropic says longer agentic tasks that search the web, run code or create files generally use more than a quick question. Claude Code adds its own drain. It sends the whole conversation with every request, so a one-line question late in a long session still pays for the history, per the Claude Code costs docs.

The numbers also move. On May 6, 2026, Anthropic said it was doubling Claude Code’s five-hour rate limits for Pro, Max, Team and seat-based Enterprise plans. Its Opus 5.5 launch post on September 22 said it was increasing five-hour usage limits on the same plans. A number from before those dates may be out of date.

The weekly limit, and Fable’s own limit on Max

Paid plans add a weekly limit that applies across all models. It is the quiet one. The five-hour limit interrupts you and comes back within hours. When the weekly limit runs out, the wait can be days.

A heavy burst can use up the week before the session window ends. Claude Code’s error reference gives a large workflow fan-out as an example.

Fable on Max: Max has a third limit, a separate weekly limit for Anthropic’s Fable models, Fable 5 and Fable 5.1, which you pick in the model picker. The Fable article says you can use up to 50% of your weekly usage limits on Fable at no extra cost. That is a cap on Fable, not a bonus. Fable draws from the same weekly limits and uses them faster than other models, and you can never use more than your weekly limit in total. At the Fable limit you can keep going on usage credits or switch to another model.

Fable on Pro: it isn’t part of the plan’s limits at all. Fable runs on usage credits from the start.

When does your Claude usage limit reset?

The session and weekly limits run on different clocks.

• Session limit: every five hours. Settings > Usage shows how much time the current session has left.
• Weekly limit: at a fixed time each week that is assigned to your account. Anthropic’s Pro and Max pages say your reset day and time “stay the same regardless of when you start using Claude or when your subscription begins.” So it is a schedule, not a seven-day clock that starts with your first message. Your next reset is in Settings > Usage.
• Fable limit on Max: resets once a week too.

Usage credits don’t change any reset time, per Anthropic’s usage credits article. An upgrade doesn’t need a reset either, because the new plan’s limits apply right away.

The reset day matters. Say your week resets Wednesday at 6:00 PM. Run out on Tuesday and you wait about a day. Run out on Thursday and you wait nearly a week.

Limit resets: now and then Anthropic gives eligible plans a limit reset. Using it puts either your five-hour limit or your weekly limit back to full, and the weekly limit still resets on its usual day. You use one with Reset for free, in Settings > Usage on the web or in Claude Desktop, or on the limit message itself. The button isn’t currently in Claude Code or the mobile app, but the reset applies there too. Unused resets expire, and the Opus 5.5 post says subscribers can now save one and use it whenever they choose.

What a Claude usage limit reached message means

The wording depends on where you are. These come from Anthropic’s error messages article for Claude chat and the Claude Code error reference.

You get a warning first. Claude shows Approaching session limit or Approaching weekly limit, and when the percentage is available, You've used X% of your weekly limit. Claude Code can show You've used 85% of your session limit · resets 3:45pm.

When it is a real limit, these are your ways through.

• Wait for the reset: the message shows the time.
• Use a limit reset: if you have one, Reset for free puts the limit back to full.
• Turn on usage credits: on paid plans, with credits on, requests switch to pay-as-you-go instead of stopping.
• Upgrade: the new plan’s limits apply right away.

Claude Code adds a few things. Since v2.1.234, a session stopped by a limit can wait and pick the task back up on its own after the reset, as long as you’re signed in with a claude.ai subscription and keep the session open, per the interactive mode docs. /rate-limit-options opens a menu for waiting, usage credits and upgrading. And if the session limit lands mid-response, Claude Code may finish the step in progress and show Usage limit reached · wrapping up. Per Anthropic’s wrap-up allowance article, it is small, capped, rolling out gradually, and the usage counts toward your weekly limit.

Some errors look like a plan limit and are not one. A length limit is about one chat getting too long, and the fix is a new conversation or a project. API rate limits are a separate system for the Claude API: Anthropic sets them per organization by usage tier, and they replenish continuously instead of resetting at fixed times, per the rate limits docs. Sign in to Claude Code with an API key instead of a subscription and there is no plan limit to hit. You pay per token.

How to see where you stand

Away from the terminal, open Settings > Usage. In Claude Code, type /usage.

• Settings > Usage on claude.ai or in Claude Desktop: on Pro, Max, Team and seat-based Enterprise plans it shows progress bars for the five-hour session and the weekly limits. Current session says how much you’ve used and how much time is left. Weekly limits says when it resets, for all models and for Fable if your plan includes it. claude.ai/settings/usage opens it.
• /usage in Claude Code: on Pro, Max, Team and Enterprise it shows plan usage bars, activity stats and a breakdown of what counted against your limits. /cost and /stats open the same screen. In the Code tab of the desktop app, click the usage ring next to the model picker.
• A status line: on Pro and Max, Claude Code can pass your status line script the five-hour and seven-day percentages and when each resets, after the first response of a session. Run /statusline and describe what you want, per the status line docs.

/status isn’t one of them. In current versions it opens a screen with your version, model, account and connectivity, even though one of Anthropic’s help articles still tells you to use it to monitor your allocation. And if ANTHROPIC_API_KEY is set, Claude Code uses that key instead of your subscription, so you pay API rates rather than using your plan’s included usage.

Our guide to how to check Claude Code usage goes through each of these, including what ccusage counts and what it doesn’t.

How to make your Claude usage limit last longer

Only tips Anthropic’s own docs give. Nothing clever. It is mostly sending less to Claude each time.

In chat, per the best practices and how limits work articles:

• Start a new chat for a new topic: longer conversations that trigger automatic context management use more of your limit.
• Combine and be specific: put related questions and the background in one message. Anthropic says planning ahead cuts the back-and-forth. On paid plans, ask Claude to search past chats instead of pasting the same context again.
• Put documents you reuse in a project: project content is cached and counts less against your limits when reused. The cache expires after a break, so the first message after one counts the content in full.
• Switch off what you don’t need: extended thinking, connected tools and web search. Anthropic notes that tools and connectors are token-intensive, and a lower effort level uses fewer tokens.
• Match the model to the task: Anthropic’s usage credits tips point to Haiku or the latest Sonnet for most tasks, at lower token cost than Opus.

In Claude Code, per the costs docs and the models, usage and limits article:

• Clear between tasks: /clear for unrelated work. Anthropic calls it “the single most effective lever for both quality and cost.” /compact keeps a summary when you need to continue.
• Choose the model for the job: Sonnet for most coding, Opus when you are stuck or the change is wide, Haiku for quick mechanical edits. /model opusplan plans on Opus and executes on Sonnet.
• Point at files, don’t paste them: pasted text stays in context for the rest of the session.
• Keep CLAUDE.md short: it rides along on every turn, and Anthropic suggests under roughly 200 lines.
• Plan before big changes: Shift+Tab cycles to plan mode, where Claude proposes an approach for your approval before it edits. That avoids expensive rework.
• Watch what runs in the background: subagents, scheduled /loop tasks and compaction send requests of their own. The /usage breakdown flags anything behind 10% or more of your recent usage and suggests a fix.
• Disable MCP servers you don’t use: /mcp shows what is configured.

My pick is the boring pair, a new chat for a new topic and /clear between tasks. They cost nothing, and they cut the history that Anthropic says grows fastest and goes out with every message.

Usage credits and when to upgrade

Usage credits are Anthropic’s pay-as-you-go option for paid plans. Turn them on in Settings > Usage (on Pro and Max, /usage-credits in Claude Code opens the page), set a monthly spend limit, add funds or turn on auto-reload, and when you hit a limit your requests switch to credits at standard API rates instead of stopping. The message changes to Now using credits or continuing with usage credits. They are billed separately from your subscription, and usage bundles save up to 30% against standard credit rates.

Max’s monthly API credits, $100 on Max 5x and $200 on Max 20x, don’t help here. Anthropic says they cover the Claude API and agents you build, not interactive Claude Code and not usage credits, and that they don’t change your usage limits. The monthly API credits article has the detail.

Upgrading is the other route. Anthropic’s own guidance in its Claude Code with Pro or Max article is to consider Max 5x if you consistently hit Pro’s limits and need more capacity for larger repositories, and Max 20x if you consistently hit Max 5x’s.

My rule of thumb is to look at which limit you hit. Hitting the session limit now and then costs you a few hours, and the habits above or usage credits can cover it. If you keep hitting the weekly limit after you’ve trimmed, that is the sign to move up. Pro to Max 5x is $80 more a month at monthly prices. For the full comparison, see Claude Max vs Pro.

On Team and seat-based Enterprise plans an admin turns usage credits on. A member without billing access can send a request to their admin from /usage-credits in Claude Code.

Keep both limits in view on a board

Settings > Usage and /usage both wait for you to look. A status line comes closest to always on, but it lives inside Claude Code. The weekly limit stings more, because it builds up all week.

Context Board is a web app at board.contextbolt.com: a live board your agents keep current. Its AI usage widget shows what is left of your plan’s current five-hour window and its week for Claude Code, and for Codex beside it, as rings that empty as you use them. They are green at 50% left or more, amber from 20% and red under 20%, each with when it resets, plus a model’s own weekly limit when your plan has one. The numbers are as each app last checked: the widget reads what Claude Code last heard from its own servers.

Agents Live Example

AI usageClaude Code · Codexnow

Claude Code

74%leftCurrent sessionresets in 3h 20m

12%leftThis weekresets Mon 9:00 AM

Codex

88%leftCurrent sessionresets in 4h 40m

79%leftThis weekresets Fri 2:00 PM

As they last checked, 2m ago

Week at 12%. The big job waits for Monday.

You add the widget from the gallery, or start from the Agents template, which already has it. A small program on your computer, set up once with one line from Connect, does the reading. Only names, states and percentages leave the computer, never what was said in a session, and more than one computer works. The widget covers Claude Code and Codex, so if you only chat in the browser, Settings > Usage stays your gauge. Running Codex too? How to check Codex usage and Claude Code vs Codex cover its side.

Context Board is free for 3 days, then $12 a month. The Claude Code usage page walks through the setup.