Claude Ultra 5-Hour Limit: What Changed from Max 20x

The Claude Ultra five-hour limit is removed only for verified accounts; the weekly cap remains. Choose Ultra when Max’s session stops interrupt long Claude Code runs and its larger weekly allowance fits your workload. Existing Ultra subscribers should check Settings > Usage before planning a long run.

What changed when Max 20x became Ultra?

The names changed, subscription prices stayed the same, and verified Ultra accounts gained freedom from the five-hour session cap. Anthropic’s October 30 announcement also says weekly limits remain unchanged. Existing subscribers do not need to switch plans manually.

As of 2026-11-02, the current Claude pricing page lists these individual plans. Prices below are monthly web prices in USD, before applicable taxes. Mobile subscription prices can differ.

Plan detail	Max	Ultra
Previous name	Max 5x	Max 20x
Monthly subscription	$100	$200
Five-hour allowance	5x Pro	No five-hour limit for verified accounts; temporary limit otherwise
Weekly allowance	5x Pro	10x Pro
Fable allowance	Up to half the weekly limit	Up to half the weekly limit
Included monthly Claude Platform API credits	$100	$200

The old Max 20x name does not describe the current weekly allowance. Budget around the published 10x Pro weekly limit. Do not infer a weekly quota increase from the removal of the session cap.

API credits are a separate entitlement with claim requirements and exclusions. They do not increase the allowance shown in your subscription usage bars.

How do you check whether your Ultra account is verified?

On an Ultra account, a missing Current session bar means the account is verified. Anthropic documents this check in What is the Ultra plan?.

1. Confirm that the account’s subscription is Ultra in Settings > Billing.
2. Open Settings > Usage.
3. Look for Current session. If the bar remains, the account still has a temporary session limit.
4. Check the weekly bars and reset time before starting expensive work.

Anthropic performs this verification automatically using account and payment history. Completing phone verification at checkout does not remove the temporary limit. The current Ultra help page gives no fixed verification deadline or manual acceleration process.

For a new account, plan around the limits actually displayed. Do not buy Ultra for an urgent overnight run on the assumption that verification will finish immediately.

This automatic check is separate from the identity-verification process, which can request government identification. Submitting identity documents is not a documented shortcut for removing Ultra’s session cap.

What still counts toward the weekly limit?

Claude Code shares the subscription allowance with Claude’s other surfaces. Moving from the terminal to web, desktop, or mobile does not create another weekly budget.

Anthropic’s usage-limit explanation separates two reset rules. A five-hour window, where applicable, starts with the first message and resets five hours later. Weekly limits reset at a fixed day and time assigned to your account. Check the displayed reset time instead of assuming the week starts with your next session.

Ultra also has an all-model weekly limit and a separate Fable weekly limit. Fable is included for up to half the weekly allowance. Together, these rules mean Fable uses part of the overall budget; it does not add another half-budget on top.

If only the Fable limit is exhausted, another model can use any remaining general allowance. If the all-model weekly limit is exhausted, switching models cannot restore included usage.

There is no reliable conversion from 10x Pro into a fixed number of coding hours or messages. Conversation size, model choice, features, and effort affect consumption. Removing session pauses lets you spend the weekly allowance faster, so a long run can bring the weekly stop forward.

Other capacity controls can still apply. The removed session cap does not increase a model’s context window or guarantee uninterrupted service.

Should you choose Max or Ultra for Claude Code?

Choose based on which limit interrupts useful work. The recurring web-price difference is $200 − $100 = $100 per month, before tax. Ultra’s published weekly allowance is 10 ÷ 5 = 2 times Max’s.

Stay on Max if your normal work fits its weekly budget and session stops are rare. Occasional paid overflow may cost less than the extra subscription charge. Measure that overflow before upgrading.

Choose Ultra if you repeatedly hit Max’s five-hour cap while still having weekly allowance left. Verification removes that particular interruption. Reconsider the upgrade if your main problem is exhausting the entire week, because Ultra still has a weekly ceiling.

For consistently heavy work, check whether twice Max’s allowance would cover the same model mix and tasks. If it would not, budget paid overflow even on Ultra. A subscription upgrade cannot turn an open-ended agent workload into a fixed-cost unlimited service.

If you also build API applications, the larger included credit may affect your decision. Count only eligible API work you expect to run before the credit expires. Someone using only interactive Claude Code should not subtract the full API-credit face value from the subscription price.

These are individual subscriptions. Teams that need organization-wide administration should assess Team or Enterprise terms separately.

Can the included API credits pay for Claude Code?

Included monthly API credits do not cover interactive Claude Code or subscription usage-credit overages. They fund eligible development work on Claude Platform. This distinction matters before treating Ultra’s $200 credit as a way around its weekly limit.

As of 2026-11-02, new subscribers can claim after seven days on an eligible plan. That is an API-credit eligibility rule, not an Ultra verification deadline. The monthly-credit help article gives this claim flow:

1. Sign in on claude.ai and open Settings > Billing.
2. In API credits, select Link organization.
3. Choose or create the Claude Console organization that should receive the credits.
4. Accept the terms and complete the link. Check the organization’s Console Billing page for the credited balance and expiry.

You need the Owner, Admin, or Billing role in the Console organization. Choose carefully: the plan links to one organization.

Credits refresh with the billing cycle, expire at its end, and do not roll over. They apply only on Claude Platform, so they cannot pay a separate gateway balance. Anthropic’s Platform credit guide explains these boundaries.

As of 2026-11-02, Anthropic’s monthly-credit FAQ confirms a specific automation exception. Included monthly API credits cover self-run claude -p and Claude Agent SDK work using an API key from the linked Console organization. When authenticated with your Claude plan instead, the same tools consume subscription allowance.

Runs started by the Claude Code GitHub Action, an IDE extension, or the desktop app remain excluded, even with -p. Do not assume that adding the flag changes the billing category.

For an interactive coding session, use purchased usage credits or purchased Console credits if you need paid continuation. Keep the included monthly credit for eligible application and agent work.

What should you do when the weekly allowance runs out?

Wait when the deadline permits, use usage credits to continue the subscription experience, or choose a separately billed API route. Check which weekly bar is exhausted first. A Fable-only stop leaves a different choice from an exhausted all-model allowance.

Wait for the displayed reset

Waiting avoids an additional usage charge. Save your work, record the unfinished task, and resume after the reset shown in Settings > Usage.

Before leaving a coding task, inspect the repository changes and record which checks have actually passed. The Claude Code checkpoint and continuation guide covers that handoff. After the reset, verify that allowance is available before restarting the task.

Enable usage credits for the current subscription

Usage credits are the most direct paid option for continuing interactive Claude Code after a plan limit. They are charged separately at standard API rates.

Follow Anthropic’s usage-credit instructions:

1. Open Settings > Usage and select Enable under Usage credits.
2. Add a payment method if required.
3. Select Adjust limit and set a monthly spending cap.
4. Select Add funds, enter the purchase amount, and confirm Purchase.
5. Leave auto-reload off unless you want automatic purchases.

Once a request uses that balance, the indicator changes to Now using credits. Work can stop at the spending cap or when the balance runs out with auto-reload off. If continuation fails, inspect those two controls before buying more capacity.

Your normal subscription reset schedule continues. Purchasing usage credits does not refill the weekly bar.

Switch to a separately billed API route

Use Console billing when you want Claude Code usage charged to a separate API organization. Anthropic confirms that purchased Console credits cover Claude Code. The included monthly promotional credit does not cover interactive sessions.

An Admin or Billing user can open the Console Billing page, select Buy credits, and fund the organization. For an existing standard Claude Code installation without custom provider configuration, exit the active session.

An ANTHROPIC_API_KEY override takes precedence over sign-in. The first command below removes it from this Bash session so the Console login can apply. Anthropic documents this in its API-key precedence guidance.

unset ANTHROPIC_API_KEY
claude auth login --console
claude auth status --text
claude

The CLI reference documents --console for Console API billing. Complete the browser login for the intended Console organization. Check authentication before sending a small request, then inspect that organization’s usage record and purchased balance. Use /status inside Claude Code to confirm the active authentication method. Authentication status alone does not prove that usable funds are available.

BetterToken is another pay-as-you-go route. Its Claude Code setup uses the Anthropic-compatible base URL https://bettertoken.ai, without /v1. Check the required model and current prices before funding its separate balance. Anthropic’s monthly credits cannot fund that balance.

How should you estimate paid overflow?

Estimate the same workload before comparing an upgrade, usage credits, or an API provider. Use the same model, currency, input/output volumes, cache assumptions, and paid-tool charges. Include applicable taxes and mandatory fees in the final budget.

For an illustrative token-only estimate, Anthropic lists claude-opus-5-5 at $4 per million input tokens and $20 per million output tokens as of 2026-11-02. The official model page confirms the ID and rates.

Assume aggregate usage across all requests of 1 million uncached input tokens and 0.2 million billable output tokens. Use standard global API pricing, with no fast mode, batch discount, cache operations, or paid tools. The estimated token charge is:

1 × $4 + 0.2 × $20 = $8 USD, before taxes.

This is a workload estimate, not a measured coding-session cost or a conversion of subscription allowance into tokens. Include repeated conversation history and tool-related tokens in the billable totals. API balances also remain subject to separate rate and spending limits.

Compare your monthly paid-overflow total with the $100 Max-to-Ultra price difference, then account for any overflow Ultra would still require. If changing models is acceptable, use the coding-task model comparison to choose a smaller trial workload. Record whether it completes the task and what it costs before changing your default.