# Claude Opus 5.5 Limit Reset: When to Spend It

> The Opus 5.5 limit reset refills your five-hour or your weekly Claude limit, not both, and expires October 22, 2026. Time it to your card and weekly reset.

- Source: https://www.aifreeapi.com/en/posts/claude-opus-message-limit-reset
- Language: en
- Published: 2025-12-14
- Updated: 2026-09-24
- Publisher: AI Free API (https://www.aifreeapi.com)

Claude Opus 5.5 arrived on September 22, 2026 with two changes for subscribers: higher five-hour session limits on Pro, Max, Team, and seat-based Enterprise plans, and a one-time **limit reset** you can save and spend when you choose ([Anthropic announcement](https://www.anthropic.com/claude-opus-5-5)). The reset is a single card in **Settings > Usage**. Depending on the card your account shows, it refills either your five-hour session limit or your weekly usage limit, not both. Your regular weekly reset still arrives on its usual day and time.

The short rule:

- **Five-hour card:** spend it when the session limit is the thing blocking you and your weekly meter still has room.
- **Weekly card:** spend it when your weekly meter is nearly full and your regular weekly reset is still days away. Spending it the day before your normal reset recovers almost nothing.
- **Either card:** use it before the expiry on the card (October 22, 2026 for most accounts), and before any downgrade or cancellation, which forfeits it.

If you are on the Free plan and saw "2 free Opus 5 messages left," that is a different thing, covered in [its own section below](#free-plan-what-free-opus-5-messages-left-means-and-is-opus-free). If you are deciding whether to move a long job to Fable 5.1, skip to [Opus 5.5 or Fable 5.1](#opus-5-5-or-fable-5-1-for-a-heavy-session).

## Which limit is actually stopping you

A Claude subscription meters four different things, and only two of them respond to the reset. Web, Desktop, mobile, and Claude Code all draw from the same pool, so a long coding session can lock you out of chat and the reverse ([Claude pricing FAQ](https://claude.com/pricing)). There is no fixed message count on any plan. How far a limit stretches depends on conversation length, files, model, features, and effort.

| Limit | Who has it | When it refills on its own | Does the limit reset help? |
| --- | --- | --- | --- |
| Five-hour session limit | Every plan, including Free | Every five hours | Yes, if your card is a five-hour card |
| Weekly usage limit (all models) | Pro, Max, Team | At a fixed weekly day and time assigned to your account | Yes, if your card is a weekly card |
| Fable share of the weekly limit | Max, Team premium seats | With the weekly limit; the share is up to 50% of it | Only indirectly, through a weekly card |
| Usage credits (prepaid, billed at API rates) | Pro and Max; Team has its own credit settings | Never; you add funds | No. A reset does not change your credit balance |

The weekly reset time does not move with your subscription date or with when you start working. The [Pro plan help page](https://support.claude.com/en/articles/8325606-what-is-the-pro-plan) and the [Max plan help page](https://support.claude.com/en/articles/11049741-what-is-the-max-plan) both say you receive the full weekly allowance each cycle and can see the next reset time in Settings > Usage. Anthropic also reserves the right to limit usage in other ways, such as model or feature caps, so a stop that matches none of these rows can happen.

Before you touch the reset, open Settings > Usage on the web or in Claude Desktop and read four things:

1. How full the current five-hour session is, and when it rolls over.
2. How full the weekly limit is, and the exact day and time of the next weekly reset.
3. The card in the **Resets** section: whether it says five-hour or weekly, and its expiry date and time.
4. On Max, how much of the Fable share you have used. On Pro or Max, your usage credits balance if you have credits turned on.

## What the Opus 5.5 limit reset does and does not do

Claude's Help Center article [What is a limit reset?](https://support.claude.com/en/articles/17007452-what-is-a-limit-reset) sets out the rules:

- It restores **either** the five-hour session limit **or** the weekly usage limit to full right away. Which one depends on the card shown in your account.
- Your weekly limit still resets on its usual day and time afterward.
- You do not have to wait until you hit a limit to use it.
- It cannot be undone.
- It does not refund extra usage you have already been billed for, and it does not change your usage credit balance.
- If you downgrade or cancel before using it, it is gone.

Some news coverage and social posts describe the reset as refilling both limits at once. The Help Center says one or the other, and user reports match that: some people hold a card that only clears the five-hour limit, others a weekly one. Anthropic has not explained how card types are assigned, so read yours rather than assuming.

The expiry comes from Anthropic's Lydia Hallie, who wrote that the reset is "available today in Settings → Usage. Apply it any time until Oct 22" ([post on X](https://x.com/lydiahallie/status/2102439437506671081)), and [Notebookcheck saw the same October 22 deadline in the app](https://www.notebookcheck.net/Claude-Opus-5-5-is-included-with-Pro-and-Free-still-misses-out.1406452.0.html). Users in Korea reported their card showing October 23 in local time. The Help Center gives no time zone, so the date and time printed on your own card is the one that counts.

### How to use the reset, including from Claude Code

1. Open Settings > Usage on the web or in Claude Desktop.
2. In the **Resets** section, click **Reset for free**.
3. Confirm by clicking **Reset for free** again.

The same button also appears on the message you see when you reach a limit. It is not available in Claude Mobile or in Claude Code in a terminal or IDE. That restriction only concerns where you press the button: limits are shared across the account, so after a reset from the browser, Claude Code and the mobile app are back to full too. If you work mainly in the terminal, keep a browser tab on Settings > Usage for the moment you need it.

## When to spend the reset: a timing rule

Because your regular weekly reset still happens on schedule, a weekly card is worth roughly the part of the weekly limit you have already used at the moment you spend it. Anthropic does not publish that as a formula, but it follows from the two rules above. Anthropic also does not publish the size of the weekly limit, so your meter percentage is the only measure you have.

**Illustration (hypothetical meter readings, not real allowances):**

| Your situation | Weekly card spent now gives you |
| --- | --- |
| Weekly meter 90% used, regular reset in 5 days | About 90% of a week's allowance, which you can use over those 5 days |
| Weekly meter 90% used, regular reset tomorrow morning | About 90% for one day, then the normal reset refills you anyway |
| Weekly meter 30% used, regular reset in 5 days | About 30%. Holding the card for a heavier week later is usually better |

That answers a common question: if you use the reset today, do you still get your full weekly reset tomorrow? Yes. But the card only bought you what you could use before tomorrow, so tomorrow's reset makes most of it redundant.

A five-hour card is worth at most one session window. It is the right card when the session limit stops you in the middle of a task while the weekly meter still shows headroom. If the weekly limit is what stopped you, a five-hour card will not get you past it.

![Decision flow: spend a five-hour card when the session limit blocks you mid-task and the weekly meter has room; spend a weekly card when the weekly meter is near full and your regular reset is two or more days away; otherwise hold it](https://www.aifreeapi.com/posts/en/claude-opus-message-limit-reset/img/reset-timing-decision.webp)

| You hold | Spend now if… | Hold if… |
| --- | --- | --- |
| Weekly card | The weekly meter is near full and your regular reset is two or more days out | Your regular reset is within about a day, or the meter is mostly empty |
| Five-hour card | You are blocked mid-task, the session rollover is hours away, and the weekly meter has room | The session rolls over soon, or the weekly limit is the real block |
| Any card, near expiry | The expiry on the card is close and you have any use for it | Never hold past the expiry; an unused card simply lapses |

On Max, a weekly card also refills the pool that your Fable share is measured against, so after using it, Fable use is again bounded by the up-to-50% share of a full weekly limit. A five-hour card does not change the weekly Fable share. On Pro, the reset does nothing for Fable spending, because Fable runs on credits and the reset does not touch credits. Both points are conclusions drawn from Anthropic's published rules, not separate policy statements.

## Opus 5.5 or Fable 5.1 for a heavy session

The two models are metered differently on each plan, and that usually decides the question before model quality does. The [Claude pricing page](https://claude.com/pricing) and the Help Center article on [Claude Fable models on your plan](https://support.claude.com/en/articles/15424964-claude-fable-models-on-your-plan) give this picture:

| Plan | Opus 5.5 | Fable 5.1 |
| --- | --- | --- |
| Free | Not included | Not included |
| Pro | Inside plan limits | Usage credits from the first request, at API rates; no one-time credit for Fable 5.1 |
| Max 5x and Max 20x | Inside plan limits | Up to 50% of the weekly limit, drawn from the same pool and used up faster than other models |
| Team | Inside plan limits | Premium seats: like Max. Standard seats: credits, like Pro |

**On Pro, stay on Opus 5.5 unless you have decided to pay for Fable.** Every Fable 5.1 request spends prepaid credits billed at standard API rates ([usage credits help page](https://support.claude.com/en/articles/12429409-manage-usage-credits-for-paid-claude-plans)). Opus 5.5 runs inside the subscription you already pay for. Current API rates per million tokens are $4 input and $20 output for Opus 5.5, and $10 input and $50 output for Fable 5.1 ([Claude API pricing](https://platform.claude.com/docs/en/about-claude/pricing)).

For a rough sense of scale, take an illustrative session of 1 million uncached input tokens and 100,000 output tokens. On Fable 5.1 credits that is 1 × $10 + 0.1 × $50 = $15. The same token volume at Opus 5.5 API rates would be 1 × $4 + 0.1 × $20 = $6, but on Pro those Opus tokens come out of your plan limits rather than your credit balance. Anthropic does not publish how a chat session maps to input, output, and cached tokens, and cache reads are far cheaper ($0.20 per million on Opus 5.5, $0.25 on Fable 5.1), so real sessions can cost much less than this example.

![Illustrative Pro session of 1M input and 100K output tokens: $15 of usage credits on Fable 5.1 versus a $6 API-rate equivalent on Opus 5.5 that counts against plan limits instead of credits](https://www.aifreeapi.com/posts/en/claude-opus-message-limit-reset/img/pro-fable-credit-example.webp)

**On Max, both models are covered, but Fable spends faster.** The Fable share is not 50% extra. It is permission to spend up to half of the same weekly limit on Fable, and Fable depletes that limit faster than other models. A Fable-heavy week can therefore leave less room for Opus 5.5 later. When you reach the Fable share, you can continue on credits or switch model. [Claude Fable 5.1 Usage Limits: Weekly Allowance, Extra Usage, and Model Switching](/en/posts/claude-fable-pro-max-weekly-limits) covers the share, credit controls, and why a conversation can switch models on its own.

**Which model should do the work?** Anthropic's own guidance is to start most work with Opus 5.5 and move to Fable 5.1 for demanding reasoning and long-horizon agentic work, or when Opus 5.5 at higher effort still falls short ([models overview](https://platform.claude.com/docs/en/about-claude/models/overview)). It also says Opus 5.5 performs at the level of Fable 5.1 on most work. That is vendor positioning, not an independent test, but combined with the metering it gives a practical order: try Opus 5.5 at a higher effort level first, and bring in Fable for the part of the job that still fails. How effort changes usage is explained in [Claude Opus 5.5 Effort: Choose a Level, Set It, and Measure the Cost](/en/posts/claude-opus-5-5-effort). If you are choosing between the models for API work rather than a subscription, see [Claude Opus 5.5 vs Fable 5.1: Which Should You Run?](/en/posts/claude-sonnet-vs-opus-vs-haiku-vs-fable)

## Claude Code: why Pro sessions now use Opus by default

If your Pro limit started running out faster in Claude Code after Opus 5.5 arrived, check your version and model. Claude Code 2.1.280 made Opus 5.5 (`claude-opus-5-5`) the default Opus model and changed the default model on Pro and Team Standard plans from Sonnet to Opus, matching Max ([Claude Code changelog](https://github.com/anthropics/claude-code/blob/main/CHANGELOG.md)). Claude's Help Center notes that Opus uses meaningfully more of your quota than Sonnet ([Models, usage, and limits in Claude Code](https://support.claude.com/en/articles/14552983-models-usage-and-limits-in-claude-code)).

What to do about it:

- Run `/model` to see the active model and switch. The conversation is kept when you switch.
- Use `/model opusplan` to plan with Opus and execute with Sonnet, which keeps Opus for the part where deeper reasoning pays off.
- Check `/effort`. An effort level you saved before per-model effort existed no longer applies to newly released models such as Opus 5.5; they start at their default until you pick a level.
- Fable 5.1 in Claude Code needs version 2.1.257 or later (the Help Center says 2.1.255, but 2.1.253 through 2.1.256 were never published), and on Pro it bills credits from the first request, the same as in the app. If it still doesn't show up, see [Fable 5.1 Not Available in Claude Code? Diagnose It by Symptom](/en/posts/claude-code-fable-5-1-not-available).

To read your limits and reset times from inside the terminal, see [Claude Code /usage: Read Limits, Resets, and Usage Monitors](/en/posts/claude-code-usage-monitor). Remember that the reset button itself is not in Claude Code.

If you had pinned Opus 5 in the Claude app, the old selection reportedly stays until you change it. Opus 5 now sits under legacy models, with no shutdown date announced.

## Free plan: what "free Opus 5 messages left" means, and is Opus free?

Opus is not part of the Free plan. The pricing page lists Opus as "No" for Free and "Yes" for Pro and Max, and Opus 5.5 does not appear for Free accounts. Free accounts do not get the limit reset card either; Anthropic offered it to subscription users.

Some Free accounts have seen something else: an unannounced Opus 5 trial labelled "3 free" in the model picker. [Notebookcheck reported on September 16, 2026](https://www.notebookcheck.net/Claude-Opus-5-starts-appearing-on-free-accounts-but-only-for-three-messages.1401190.0.html) that the count drops with each message, from "2 free Opus 5 messages left" to "1 free Opus 5 message left," and that after the last one Opus 5 is gone from that account. It does **not** come back when the five-hour session limit refreshes. Anthropic has not announced the trial or said how accounts are chosen, and most Free users still see an upgrade prompt next to Opus. Whether it ever returns is unknown, so treat those three messages as a one-off.

What Free users can count on is the regular five-hour session limit on Sonnet and Haiku. Getting Opus 5.5 regularly requires a paid plan; Pro is $20 per month in the US, or $17 per month on an annual plan ($200 billed up front). Whether Pro's weekly limit is enough for your workload, or whether Max makes sense, is covered in [Claude Pro vs. Max Usage Limits: The Weekly Cap Still Applies](/en/posts/claude-code-pricing-pro-vs-max).

## What Anthropic has not published

- **How much the five-hour limit went up.** The announcement confirms an increase but gives no figure. Percentages circulating online have no Anthropic source.
- **Whether weekly limits changed** with the Opus 5.5 launch. The announcement mentions five-hour limits only.
- **How card types are assigned**, and why some accounts hold a five-hour card while others hold a weekly one.
- **Whether the Free Opus trial will recur.**

For API model IDs, prices, and migration notes rather than subscription limits, see [Claude Opus 5.5: API Price, Model ID, and Migration Checks](/en/posts/claude-opus-5-5).

## Quick answers

### Does the Opus 5.5 reset refill both the five-hour and weekly limits?

No. Each card restores one of them, as shown on the card in Settings > Usage. Reports claiming both refill do not match the Help Center.

### If I use the reset today, do I still get my normal weekly reset?

Yes. The weekly limit resets on its usual day and time regardless. That is why a weekly card spent shortly before your normal reset gives you very little.

### Can I use the reset in Claude Code or on my phone?

Not from there. Press **Reset for free** on the web or in Claude Desktop. The refilled limits then apply in Claude Code and the mobile app, because the pool is shared.

### Does the reset give me Fable 5.1 usage on Pro?

No. On Pro, Fable 5.1 runs on usage credits from the first request, and a reset does not change your credit balance.

### When does the Opus 5.5 limit reset expire?

October 22, 2026 for most accounts, according to Anthropic staff and the in-app deadline; some users in Korea saw October 23 local time. Go by the date and time on your own card.