Claude Code Usage Limits, Explained: The 5-Hour Window, the Weekly Cap, and What Changed in October 2026

Updated October 27, 2026 · first published September 23, 2026 · by the AQ team

Claude Code runs on your Claude subscription, and as of October 2026 that subscription carries two stacked usage limits: a session limit that resets every five hours, and a weekly limit that resets at a fixed time Anthropic assigns to your account. Max plans add a third meter, a separate weekly limit for Fable models. All of them are shared across Claude Code, the claude.ai chat, the desktop app and IDE extensions, so one pool pays for everything. The fastest way to see your remaining budget is to type /usage inside a Claude Code session. Since the September 2026 version of this guide, Anthropic has documented a wrap-up allowance that lets a response in progress finish after the five-hour limit lands, moved usage credits into Claude Code itself with /usage-credits, and listed new API prices for Sonnet 5.5 and Opus 5.5. This guide covers how each limit works, what drains them, how to check, and what to do at the limit.

How does the 5-hour session limit work?

A session is a rolling five-hour usage window: it opens when you send your first message, everything you do for the next five hours draws from that window's allowance, and the session limit resets every five hours. Two consequences follow. The reset time depends on when you started, not on the clock: begin at 8:07am and the window resets around 1:07pm. And a session is a usage budget, not a time budget: light use for five hours may barely dent it, while one heavy agentic run can exhaust it in under an hour. Claude Code shows warning messages as you approach any limit.

Anthropic sizes plans relative to Pro rather than publishing token counts: Max 5x includes five times the Pro plan's per-session allowance and Max 20x twenty times. Team premium seats get five times a standard seat.

How does the weekly limit work?

The weekly limit is the ceiling above the sessions: a budget across all models that resets once a week at a fixed day and time assigned to your account, visible in /usage or at Settings then Usage on claude.ai. You receive your full weekly allowance each cycle.

The weekly numbers moved in 2026. A limited-time promotion raised weekly Claude Code limits by 50 percent for Pro, Max, Team and seat-based Enterprise plans from May 13 through September 13, 2026 (five-hour limits were not part of it). Since September 14, 2026 weekly limits sit permanently 25 percent above their pre-promotion level. Measured against the promotional level that is about 17 percent less weekly headroom, so if your weekly budget felt tighter from mid-September 2026, that change is the reason.

The wrap-up allowance: finishing a step after the limit lands

Reaching the five-hour limit used to stop Claude Code in the middle of whatever it was doing. Anthropic now documents a wrap-up allowance: when the limit lands during a response, Claude Code may keep working briefly so it can finish the step in progress, and the session shows "Usage limit reached · wrapping up." The allowance is small and capped, it varies by plan, and it may not be enough to finish the whole task. Its usage counts toward your weekly limit and does not raise it.

The rules to know: it applies only to a response already in progress, never to a new message you send after reaching the limit. On Pro it applies up to once per weekly usage period; on Max and Team premium seats it can apply each time you reach a five-hour limit, within your weekly limit. It needs Claude Code 2.1.277 or later and a claude.ai subscription login (not an API key or a third-party cloud provider), and it is rolling out gradually, so it may not be on your account yet. If you have usage credits enabled, the included wrap-up allowance is used first and the response then continues on credits at standard rates.

Fable models have their own weekly limit

Fable 5 and Fable 5.1 are available on Pro, Max, Team and Enterprise, but how they are metered depends on the plan. On Max, Team premium seats and premium seat-based Enterprise seats, you can use up to 50 percent of your weekly usage limit on Fable models at no extra cost; that Fable usage also draws from your regular weekly limit, and it depletes it faster than other models. Once you hit the Fable limit you can switch to another model or continue on usage credits. The 50 percent cap does not add to your total weekly limit.

On Pro and Team standard seats, Fable models are not part of the plan's included usage and run on usage credits from the first message. In Claude Code, Fable 5 needs version 2.1.170 or later and Fable 5.1 needs 2.1.255 or later. /usage shows the Fable meter separately on plans that include it.

What actually draws down the limit?

Usage is metered in tokens processed, and per Anthropic's help center it scales with message length, attachment size, current conversation length, tool usage, model choice, effort level and multi-step tasks such as running code or creating files. In Claude Code the expensive part is rarely your prompt: every turn resends the conversation so far, your CLAUDE.md project context, and every file Claude has read this session, then adds the model's output, tool results, and thinking on top. Prompt caching re-reads that history at the cached rate, but a one-line question in a session that has been open all day still draws usage for the whole conversation, and the first message after a break longer than the cache lifetime reprocesses the full context. A fresh session per task remains the single cheapest habit.

Model choice is the other big lever. Anthropic's own Claude Code documentation says Opus uses meaningfully more of your quota and recommends Sonnet for the large majority of coding work, with Haiku as the fastest and cheapest option. Fable models deplete the weekly limit faster still. Subagents, scheduled tasks and agent teams all send their own requests on top of the main conversation, and the /usage breakdown (below) shows how much of your recent usage each one accounts for.

What each plan includes, as of October 2026

Prices from claude.com/pricing and Anthropic's help center, read October 27, 2026. The multipliers describe per-session capacity relative to Pro; Anthropic does not publish absolute token counts per plan.

Usage credits and monthly API credits are different things

Two kinds of credit appear in a Claude account, and readers mix them up constantly. Usage credits are a prepaid balance you enable on Pro, Max, Team or Enterprise so that work continues after your included limits run out, billed at standard API rates; they work in Claude Code and the Claude apps. Monthly API credits are the $100 (Max 5x) or $200 (Max 20x) a month, or the pooled Team balance capped at $500, that Anthropic grants for the Claude API, the Console Playground, Claude Managed Agents and the Agent SDK. Anthropic's help center is explicit that monthly API credits cannot be used in Claude Code or the Claude apps, including for usage credits, and they do not change your plan's usage limits. They refresh each billing cycle, do not roll over, and are claimed by linking a Console organization from Settings then Billing after seven days on the plan.

How do I check how much usage I have left?

Inside a Claude Code session, type /usage. It draws progress bars for your current five-hour session and your weekly limit, with the reset time for each, plus the Fable meter on plans that have one. The Session block at the top shows API token counts and a dollar estimate, which matters only if you bill through an API key; subscribers can ignore the cost figure and read the plan usage bars. On claude.ai, Settings then Usage shows the same meters. The separate /cost command tracks dollar spend for API-key billing.

On a Pro, Max, Team or Enterprise plan, /usage also breaks down what counts against your limits: recent usage attributed to skills, subagents, plugins and individual MCP servers, each as a percentage of the total, plus recommendations such as "Trim CLAUDE.md" or an /effort change, each prefixed with the share of recent usage it applies to. Press d or w to switch between the last 24 hours and the last 7 days; the figures come from local session history on that machine. When the footer shows a hint such as "f to fix 2 of 3", pressing f has Claude Code apply the recommendations it can (Claude Code 2.1.305 or later). While usage credits are on, /usage also shows a usage-credits row with this month's spend against your monthly spend limit.

One note that is easy to misread: when Claude Code cannot fetch your plan limits, usually because the usage endpoint itself is rate limited, /usage shows the last bars it loaded on that machine within the past 60 minutes with a "Showing last-known usage" note, and a message like "rate limited until 3:14am" means only that the meter cannot refresh until then. It is not a usage limit on your work; press r after that time for fresh data.

What happens when you hit a limit?

Claude Code tells you when the limit resets and, if a wrap-up allowance applies, finishes the step in progress first. Nothing is lost: the conversation history and your working tree stay intact. On Claude Code 2.1.234 or later, an interactive session signed in with a claude.ai subscription waits in place and continues the interrupted task on its own after the reset ("Usage limit reached · limit resets 3:45pm", "Continuing automatically at 3:45pm · esc to cancel"), re-arming at most twice in a row; /rate-limit-options turns that off or back on.

From there you have four options. Wait for the reset (at most five hours for a session limit). Switch to a cheaper model with /model: after a model-specific message such as "You've hit your Opus limit", a model outside that family keeps you working, but a session or weekly limit covers every model, so switching does not restore access. Enable usage credits: run /usage-credits inside Claude Code (it needs a /login with your claude.ai subscription, not an API key), which opens Settings then Usage for Pro and Max subscribers, where you turn credits on, add a payment method, set a monthly spend limit (or no limit at all), prepay a balance and optionally turn on auto-reload. Team and Enterprise members with billing access land in their organization's usage settings; members without it confirm a request that goes to their admins. Once your included limit is used up, requests move to credits without cutting you off, and the notification changes to "Now using credits"; when you reach your spend limit with credits still available, Claude Code prompts you to raise or remove the limit without leaving the CLI. Credits bill at standard API rates, do not change your reset schedule, and generally do not expire (in certain jurisdictions, including Japan, credits bought on or after September 10, 2026 expire six months after purchase). If you subscribed through a Claude mobile app, you can only enable and purchase usage credits on the web. Or upgrade the plan outright, Pro to Max 5x or Max 5x to Max 20x.

Scheduling work around the windows

Teams that live inside these limits converge on the same habits:

• Start long runs at the top of a window. A five-hour window you open at 9am is gone by 2pm regardless of what you did with it. Kicking off the big refactor right after a reset, instead of 20 minutes before one, is free capacity.
• One task per session, one session per worktree. Context accumulation is the silent drain. A fresh session in an isolated git worktree starts from zero context instead of paying for the last task's leftovers on every turn.
• Parallelize tasks, not marathons. Three focused sessions on three worktrees finish more work per token than one nine-hour session, because none of them re-reads a bloated history. Running multiple Claude Code sessions in parallel covers the mechanics; parallel sessions do all draw from the same account pool at once.
• Keep the machine running. Limits are account-scoped, not machine-scoped, so a server buys no extra tokens. What it buys is that the usage you have lands in finished work: a run that survives a closed laptop uses the window you paid for, and the automatic wait-and-continue only helps if the session is still open at the reset.
• Split people, and sometimes accounts. Each teammate's subscription is its own pool, so a team's real capacity is per person. Some individuals hold more than one subscription they own; running multiple Claude Code accounts explains the setup and when it is worth it. Sharing one login between people is prohibited by Anthropic's consumer terms.

When does API billing beat a subscription?

As of October 2026, claude.com/pricing lists Claude Sonnet 5.5 at $2 per million input tokens and $10 per million output tokens, Claude Opus 5.5 at $4 and $20, Claude Haiku 5.5 at $0.10 and $0.50 for prompts up to 100K tokens, and Fable 5.1 at $10 and $50. A subscription is a bulk discount with a ceiling; the API is metered with no ceiling. Steady daily interactive use almost always favors the subscription; the API wins for spiky automation such as CI jobs, batch refactors, or an occasional very heavy week where the cap would stall you. The pragmatic middle path is usage credits, which bill overflow at standard API rates: one month with them enabled tells you what uncapped API billing would cost. Remember that Max and Team monthly API credits do apply to API-key usage, including claude -p runs with a linked Console key, so a Max subscriber who also automates through the API may already have that spend covered. Coding agent token costs for teams goes deeper on budgeting this across a team.

Where AQ fits

AQ is the multiplayer coding harness where engineering teams run AI coding agents like Claude Code and Codex together: shared live terminals, a code editor, and app previews, in your own cloud. AQ does not change your Claude limits and does not resell model access: each person connects their own Claude, OpenAI or Cursor subscription, and AQ never marks up usage on the accounts your team connects. What it changes is how much of your allowance becomes finished work. Sessions run on a machine dedicated to your team, so the run you start at the top of a five-hour window keeps working after your laptop closes and is still there, resumable from any device, when the window resets. Each work item gets its own isolated git worktree, which makes fresh-session-per-task the default rather than a discipline, and teammates can open the same live session. A person can connect more than one account of the same CLI, and a thread whose account reaches its plan window can continue on another connected account that still has headroom, the same situation running multiple Claude Code accounts covers by hand. The Free plan is a personal sandbox AQ creates for you with nothing to install and no time limit; the AQ Pro plan is $50 per user per month in early access (standard $200), locked for your first 12 months, and each seat includes $10 of AQ credits a month for AQ's own models.