Claude Pro and Max Limits After the Permanent 25% Claude Code Increase

Claude Pro limits: what changed on September 14?

The permanent increase applies to Claude Code weekly limits on Pro, Max, Team, and seat-based Enterprise plans. The key comparison is with the allowance before the promotion, rather than the larger allowance available during it.

The official weekly-limit promotion page states that weekly limits in Claude Code are now 25% higher than before the promotion. It does not publish an absolute weekly token allowance or a guaranteed number of coding tasks.

An index makes the change easier to understand:

These numbers express relative capacity; they are not token or message quotas.

Why Claude Code can feel more limited now

The temporary promotion ended September 13 at 11:59 PM PT. Moving from an allowance of 150 to 125 leaves 16⅔% less weekly capacity than during the promotion.

That percentage is arithmetic derived from the announced allowances, not a separately announced reduction. Both descriptions can therefore be true: Claude Code has a permanent increase over its original baseline, and developers have less weekly room than they had during the promotional period.

The promotional boost applied automatically. Free plans and consumption-based Enterprise seats were excluded.

Weekly limits and five-hour limits answer different questions

A weekly limit governs usage across a longer period. A five-hour limit governs the shorter session window. Increasing one does not automatically increase the other.

The weekly promotion covered Claude Code across its CLI, IDE extensions, desktop, and web interfaces. It did not increase five-hour limits or the limits for Claude chat and Claude Cowork.

The separate May five-hour increase

On May 6, Anthropic separately doubled Claude Code’s five-hour rate limits for Pro, Max, Team, and seat-based Enterprise plans. It also removed the peak-hours limit reduction for Claude Code on Pro and Max.

That earlier change matters when interpreting the September announcement. The permanent weekly increase does not represent another doubling of five-hour capacity.

If you encounter a limit while working, first identify which window is exhausted. A larger weekly allowance does not guarantee that a concentrated coding session can continue through its five-hour limit. Our Claude Code rate limits guide provides a related reference for understanding these limits.

Pro, Max, and shared usage: avoid fixed quota assumptions

Pro and Max include Claude and Claude Code under one subscription. Activity in both counts against shared usage limits, as documented in the subscription guide dated August 19, 2026.

For developers, the practical implication is straightforward: consider chat activity when assessing how much subscription usage remains for coding. Do not budget the two interfaces as independent pools.

How many messages or coding tasks does Pro include?

The evidence available for this article does not establish a universal daily message allowance or weekly coding-task quota. The promotion announcement gives percentage increases without publishing absolute weekly quantities.

Pro’s capacity can depend on message length, attachments, conversation length, model, and feature. A fixed “tasks per week” estimate would hide those differences and suggest a guarantee that the available evidence does not support.

Use the allowance and reset information shown for your account when planning work. Treat any estimate based on a particular prompt or workflow as an estimate for that workload.

What do Max 5x and Max 20x mean?

The available Max plan description defines these multipliers as five or twenty times Pro’s per-session usage allowance. Its exact wording as of September 16 was not independently verified, so those labels should not be converted into a claimed historical weekly token quota.

For a purchasing decision, distinguish the session allowance from the weekly allowance. Do not assume “Max 20x” guarantees twenty times as many completed coding tasks: the work itself varies, and the published weekly promotion gives no fixed task count.

Pro’s $20 monthly price is supported by its dated launch announcement. This article leaves out exact Max subscription prices because the supplied evidence does not independently establish them at the publication cutoff.

Check usage and confirm which account is paying

Before upgrading or buying credits, inspect your usage and authentication. These checks address two different problems: exhausting the subscription allowance and accidentally using API billing.

Check usage in Claude Code

The documented CLI command for viewing the increased limits is:

/usage

Review the usage and reset information available for your account. The evidence does not establish one universal reset schedule for every Pro and Max account, so use the account’s displayed information when deciding whether to wait.

If you need help interpreting a blocked session, see the usage-limit-reached guide.

Switch to subscription authentication

The August 19 guide documents this command for switching from Claude Console pay-as-you-go authentication to the subscription:

/login

An existing ANTHROPIC_API_KEY environment variable takes precedence over subscription authentication and results in API charges. A paid Pro or Max subscription therefore does not, by itself, establish how a particular Claude Code session is billed.

For account-selection problems, the same guide documents /logout, claude update, and claude. Check the account you select when signing in.

API billing is a separate option for these workflows. The Claude Code API key setup guide covers that route, including use through an independent API gateway such as AI Prime Tech.

Claude extra usage: costs, credits, and when they are worth it

When included usage runs out, the August 19 subscription guide lists several options: wait for a reset, upgrade, enable usage credits, or purchase Claude Console API credits. These are different choices, so first decide whether you need occasional continuation or regularly need a larger subscription allowance.

“Extra usage” is now “usage credits” in the CLI

Claude Code v2.1.144, released May 19, renamed extra usage to usage credits in CLI copy. The command became:

/usage-credits

The old command still works:

/extra-usage

Readers searching for “Claude extra usage” may therefore encounter “usage credits” in Claude Code. The terminology change does not itself establish a change in the included subscription allowance.

What the dated bundle guide establishes

The usage-bundle guide dated May 18 documents these options for Pro and Max subscribers:

Bundles apply after included limits are exhausted. The documented purchase path is Settings → Usage → “Buy usage.”

The monthly discounted-purchase cap was $2,000 in credit value. That figure describes the cap on discounted credit purchases; it should not be treated as a personal spending budget.

The supplied usage-credit billing guide describes continuation at standard API rates, separately charged from the subscription, with prepaid funds and a monthly spending cap. However, its exact September 16 wording was not verified. Confirm the applicable rates and spending controls before enabling paid continuation; this article does not backdate unverified button labels or auto-reload rules.

Is extra usage for paid Claude plans worth it?

Extra usage is worth considering when a limit blocks valuable work and you can set an acceptable budget for continuing. Compare the credit purchase cost with the value of finishing now and the option of waiting for the displayed reset.

For recurring interruptions, assess whether a larger subscription allowance better fits the workload. For occasional interruptions, credits may be a more targeted purchase. Neither choice guarantees a fixed number of tasks, and a bundle discount alone does not establish that the purchase is economical for your work.

A practical decision sequence is:

1. Check which usage window is exhausted and when it resets.
2. Confirm whether the session uses subscription or API authentication.
3. Compare waiting, upgrading, and buying credits.
4. Verify the applicable rates and spending controls before proceeding.

Key takeaways

• Claude Code weekly limits became permanently 25% higher than their pre-promotion baseline on September 14, 2026, for eligible paid plans.
• The permanent allowance is 16⅔% below the temporary promotional allowance, which explains why weekly capacity can feel lower after September 13.
• The weekly promotion did not increase five-hour limits, Claude chat limits, or Claude Cowork limits.
• Pro and Max share usage between Claude and Claude Code; the evidence does not establish a universal message or coding-task quota.
• Use /usage to inspect limits, and check authentication before paying: ANTHROPIC_API_KEY takes precedence over subscription login.

FAQ

Does the permanent 25% increase raise Claude Pro’s five-hour limits?

No. The September change increases Claude Code weekly limits relative to their pre-promotion level and does not increase five-hour limits. Claude Code’s five-hour rate limits had separately doubled on May 6 for the eligible paid plans.

Do Claude chat and Claude Code share Pro and Max limits?

Yes. The August 19 subscription guide states that activity in Claude and Claude Code counts against shared usage limits on Pro and Max. The weekly promotion increased Claude Code limits without increasing Claude chat limits.

How much does Claude extra usage cost?

The May 18 bundle guide documents $50 in credit value for $45, $250 for $200, and $1,000 for $700. These bundles apply after included limits are exhausted, with a monthly discounted-purchase cap of $2,000 in credit value. Exact September 16 wording for the broader usage-credit billing controls was not verified.

Why am I receiving API charges while subscribed to Pro or Max?

An existing ANTHROPIC_API_KEY environment variable takes precedence over subscription authentication in Claude Code and results in API charges. The documented command for switching from Claude Console pay-as-you-go authentication to a subscription is /login. Check the environment variable and selected account when investigating unexpected charges.