Kingy AI
Subscribe, 2.1M subscribersWork with Kingy
AI News

Claude Max 20x Is Now Ultra: What Changed for Paid Plans

April 15, 2026 Blog
Claude pricing plans and usage billing editorial image

Anthropic has renamed Claude Max 20x Ultra. Its help page says existing subscribers move automatically, without a price change.

The current individual lineup is Free, Pro, Max and Ultra. Team and Enterprise retain their names on the pricing page; this article concerns individual plans. The top tier costs $200 a month on the web; Max costs $100. The more consequential distinction is the period attached to the advertised allowance: the old 20x described a session, while Ultra’s 10x describes a week.

We checked Anthropic’s pricing and support documentation on October 30, 2026. We did not test account migration, inspect a subscriber’s invoice or measure plan capacity. The documents establish the published terms; they do not establish the exact rename launch date or when any particular account completes the move.

Claude’s old and new plan names

Previous individual plan Current comparison How to read it
Free Free The no-cost option.
Pro Pro The paid baseline.
Max 5x Max The $100 tier is the closest counterpart.
Max 20x Ultra Anthropic explicitly documents this rename.

The Max 5x mapping follows the price and advertised tier; we did not find an equally explicit migration statement for that group. Anthropic’s current Max page describes the present offer. Its German-language Max page still describes the earlier two tiers: $100 for 5x Pro usage per session, or $200 for 20x. It says sessions reset every five hours and a separate weekly cap applies, without supplying a numerical Pro-relative weekly allocation.

That localization is useful evidence of the earlier wording, but its old plan names make it unsuitable as a guide to current English terms. A screenshot or search result that says Max 20x needs the same caution: check its date and the period attached to the number.

Current Claude plan prices and allowances

These are Anthropic’s advertised US web prices, excluding applicable taxes. Mobile-store prices can differ. The allowances are vendor comparisons with Pro, rather than measured results. Sources: pricing, Pro, Max and Ultra.

Plan US web price Five-hour allowance Weekly allowance Fable access Monthly Platform API credits
Free $0 Limited Limited No None
Pro $20 monthly, or $200 annually Pro baseline Pro baseline Paid usage credits None
Max $100 monthly 5x Pro 5x Pro Within 50% of weekly pool $100; eligibility and claim required
Ultra $200 monthly No five-hour cap after account verification 10x Pro Within 50% of weekly pool $200; eligibility and claim required

Pro’s $200 annual payment works out to about $16.67 per month: $200 divided by 12. That saves $40 against twelve $20 monthly payments, before tax. The pricing card rounds the monthly equivalent to $17; a buyer cannot pay $17 each month on a month-to-month subscription. Max and Ultra are currently monthly-only plans.

The annual calculation assumes you would otherwise keep paying for all twelve months. Someone who needs a paid subscription for only part of the year should compare the total payments for that shorter period, rather than treating an annual discount as an automatic saving.

Why old 20x and new 10x are not a direct comparison

The historical session multiple and current weekly multiple measure different periods. Dividing 10 by 20 therefore cannot establish a 50% reduction. The material we checked also cannot support a promise that former Max 20x subscribers retain exactly their previous weekly capacity. A clear historical weekly figure or a matched before-and-after workload would be needed for a narrower conclusion.

The present Max offer provides 5x Pro usage for both the session and the week, according to its help page. Ultra’s weekly multiple is twice that, at twice the monthly price. This is arithmetic on advertised allowances; it does not mean twice the answers, twice the coding hours or faster inference.

Anthropic’s usage guide says model choice, conversation complexity, features and effort affect consumption. Claude’s product surfaces share usage limits. Where a five-hour cap applies, the period starts with the first message and resets five hours later. Weekly resets follow a fixed schedule assigned to the account. Verified Ultra accounts have no five-hour cap, but weekly limits remain, and Anthropic may impose other capacity controls. A usage allowance is also different from the context window, which concerns the information a model can handle within a conversation.

For a buyer, the useful distinction is which meter interrupts the work. An occasional concentrated session and a week of steady use can produce different constraints. Track a representative week before upgrading: record what you were doing when you hit a limit and which meter was exhausted. The plan name alone cannot answer that question.

Check Ultra verification in Settings > Usage

Anthropic verifies accounts automatically using account and payment history. An unverified Ultra account retains a temporary restriction. In Settings > Usage, absence of a Current session bar indicates verification. Phone verification at checkout is separate. The Ultra article gives no guaranteed waiting period or manual self-verification procedure.

Before choosing the top tier primarily for long sessions, inspect the account you intend to use. Public policy cannot establish that account’s status. If the account state or renewal details conflict with the published terms, keep the relevant screen and billing details for a support request.

Fable shares the weekly pool

According to Anthropic’s Fable guide, Fable 5 and 5.1 are included on Max and Ultra, drawing from the regular weekly allowance and consuming it faster than other models. Fable can use up to half that allowance; it does not add another 50% on top. After the Fable cap, subscribers can switch models or buy usage credits. Pro requires usage credits for Fable from the start; Free has no access.

If Fable drives your workload, watch both the overall weekly meter and the model-specific meter. Reaching the latter need not mean the entire weekly pool has been spent. Avoid budgeting as though every unit advertised for the plan can be used with your preferred model.

API credits and paid usage credits are different

Anthropic’s API-credit guide requires an active plan in good standing; new subscribers wait seven days. Claim on web Billing with a Console Owner, Admin or Billing role. Link one organization; each organization can receive credits from only one plan. Changing it requires support. No card is required to claim.

Credits refresh each billing cycle, expire without rollover and precede purchased credits. Organization API keys share the balance. Eligible uses include Messages/Batch APIs, Playground, Managed Agents and Agent SDK. Self-run claude -p or SDK use qualifies with the linked organization’s API key; plan-authenticated use draws on subscription limits. Interactive Claude Code, extra use in Claude/Cowork, and Bedrock, Vertex AI or Foundry are excluded. GitHub Action, IDE-extension and desktop-app Claude Code runs are excluded even with -p.

For a claimed credit, Max-to-Ultra upgrades provide a prorated adjustment, then the higher amount in subsequent cycles. Choose the receiving organization carefully before connecting it.

The subscription’s separate usage-credit system pays for extra use in Claude, Claude Code and Cowork at standard API rates. Monthly Platform API credits cannot pay those overages. Usage credits create additional charges; monthly spending caps and optional auto-reload are available. Anthropic’s credit terms say credits have no cash value and auto-reload authorizes recurring payments.

A developer with eligible API work may find value in the included amount. A subscriber who only chats on the web should not subtract it from the subscription price. Enabling paid extra chat can create charges even if the API-credit balance is untouched. Before enabling either kind of paid continuation, decide which work should continue and what you are willing to spend.

API credits also predate this naming change: Anthropic’s October 7, 2026 release note announced them for Max and Team. A current pricing-card badge does not establish that the benefit launched with Ultra.

Which plan fits your usage?

Pro is a reasonable starting point if its baseline allowance covers the work. The annual option deserves consideration when you expect to keep the subscription for a full year and can make the upfront payment. If a particular model is essential, include any separate model usage in your budget.

Max is the comparison to make when Pro regularly interrupts work. Review how often that happens and what the interruption costs you before committing to the higher monthly bill. An included credit should influence the choice only when you can claim and use it for work you would otherwise pay for.

Ultra deserves consideration when the lower tier leaves a clear capacity problem. A researcher concentrating work into long sessions may have a different need from a developer exhausting a weekly allocation. Use the meters and the account check above to identify your constraint. If neither is a problem, the additional $100 monthly cost needs a stronger reason than the new name.