Claude Max vs Claude Ultra: Complete Comparison and Report on Price, Usage Limits, 5-Hour Window, and API Credits

Claude Max and Claude Ultra are the two individual Claude subscriptions that sit above Pro. Max costs $100 a month and gives five times the Pro plan's usage, both inside every five-hour session and across the week. Ultra costs $200 a month and is the new name, adopted by Anthropic on October 30, 2026, for the plan previously sold as Max 20x. With the rename, Anthropic described Ultra purely by its weekly allowance, set at ten times Pro, and removed the five-hour session limit for verified Ultra accounts. Existing Max 20x subscribers were moved to Ultra automatically at the same price, and the $100 plan formerly called Max 5x is now simply Max.
The two plans cost exactly the same per unit of usage. Max buys 5x Pro's weekly usage for $100 and Ultra buys 10x for $200, so both work out to $20 a month for each Pro-week of capacity, the same rate as a monthly Pro subscription. What the extra $100 buys on Ultra is twice the weekly allowance, no five-hour window once the account is verified, twice the absolute Fable budget (both plans cap Fable at half of the weekly limit) and $200 instead of $100 a month in Claude API credits. Every other feature listed on Anthropic's pricing page for Max is included in Ultra.
Max is the better choice for users whose work fits comfortably inside 5x Pro per week and who rarely hit the five-hour session limit, including people who use Claude Code in short, frequent bursts. Ultra is the better choice for developers who run long Claude Code or agent sessions that stall at the five-hour cap, for heavy Fable 5.1 users, and for anyone who regularly spends more than $100 a month on usage credits on top of Max. All prices and plan terms in this comparison were verified on Anthropic's official pricing page and Claude Help Center on November 7, 2026.
··········
CLAUDE MAX AND CLAUDE ULTRA: HOW THE TWO PLANS REACHED THEIR CURRENT FORM.
Until the end of October 2026, Anthropic sold one Max plan in two sizes, Max 5x at $100 and Max 20x at $200. The multipliers in those names described the five-hour session allowance, which caused confusion because the weekly allowance of Max 20x was 10x Pro, half of what the name suggested. The October 30 change replaced the naming with a structure based on the weekly limit and, at the same time, turned the $200 tier into the only individual plan without a five-hour window.
Several other changes in October and November affected both plans. Monthly API credits for Max arrived on October 7, the Claude Duet research preview opened to Pro and Max on October 21, and the Personal Finance beta launched on November 5 for Pro, Max and Ultra in the United States.
........
Date (2026) | Change | Max | Ultra |
October 7 | Monthly Claude API credits introduced for Max and Team | $100 a month | $200 a month, as listed for Ultra after October 30 |
October 21 | Claude Duet research preview opens | Included | Included (Duet help article lists Ultra) |
October 30 | Max 5x renamed Max, Max 20x renamed Ultra | Price unchanged at $100 | Price unchanged at $200, 5-hour limit removed for verified accounts |
November 5 | Personal Finance beta, US only | Included | Included |
........
The rename did not change the weekly capacity of either plan. The Ultra help article states that the plan provides 10x Pro's usage every week, and Anthropic's announcement said weekly limits stayed the same, so a former Max 20x subscriber has the same weekly budget as before and loses only the five-hour cap once verified. Our report on the Max 20x rename and the end of Ultra's 5-hour limit covers the announcement itself.
··········
CLAUDE MAX VS CLAUDE ULTRA PRICE AND BILLING TERMS.
Both plans are sold only as monthly subscriptions. Anthropic's pricing page shows no annual option for Max or Ultra, and the pricing FAQ confirms that both are billed monthly. The listed prices apply to web subscriptions; the help articles for both plans state that pricing through mobile app stores may differ, and that prices exclude applicable tax.
........
Billing item | Claude Max | Claude Ultra |
Monthly price (web) | $100 | $200 |
Annual billing | Not offered | Not offered |
Cost over 12 months | $1,200 | $2,400 |
Mobile app price | May differ by app platform | May differ by app platform |
Upgrade from Max | Not applicable | Prorated for the rest of the billing cycle |
Upgrade from annual Pro | Unused Pro balance above the Max price is credited | Unused Pro balance above the Ultra price is credited |
Google Play upgrade | Full new price charged, unused value becomes extra days | Full new price charged, unused value becomes extra days |
Standing discounts | None for individual plans | None for individual plans |
After cancellation | Moves to Free at the end of the period | Moves to Free at the end of the period |
........
The upgrade rules are the same for the two plans. A subscriber on annual Pro who moves to Max or Ultra receives a credit when the unused Pro balance exceeds the new plan's price, but the billing address must match the previous subscription; if it has changed, the subscriber has to cancel Pro, wait for the end date and then subscribe again. Moving from Max to Ultra is charged on a prorated basis for the remainder of the billing cycle. Anthropic's help center states that there are no standing discounts on Pro, Max or Ultra and that support cannot issue coupons; discounted pricing for nonprofits applies only to Team and Enterprise. Cancellation must happen at least 24 hours before the renewal date, and both plans fall back to Free when the paid period ends.
··········
USAGE LIMITS ON CLAUDE MAX AND CLAUDE ULTRA.
Anthropic expresses both plans as multiples of Pro and does not publish absolute message, token or hour counts for any individual plan. The Help Center lists the factors that consume usage: message length, attachment size, conversation length, tool use such as Research and web search, the selected model, the effort level, Artifacts and multi-step tasks such as running code or browsing. Usage in the Claude web, desktop and mobile apps and in Claude Code draws from the same pool.
........
Limit | Claude Max | Claude Ultra |
Five-hour session limit | 5x Pro | None for verified accounts, temporary limit until verification |
Weekly limit, all models | 5x Pro | 10x Pro |
Separate weekly Fable limit | Yes, up to 50% of the weekly limit | Yes, up to 50% of the weekly limit |
Shared with Claude Code | Yes | Yes |
Usage credits after a limit | Available (standard API rates) | Available (standard API rates) |
Limit resets | Can restore 5-hour or weekly limits | Can restore weekly limits |
Additional caps at Anthropic's discretion | Possible (weekly, monthly, model or feature limits) | Possible (weekly, monthly, model or feature limits) |
........
The difference in structure is larger than the 2x weekly figure suggests. A Max subscriber works against two ceilings, a five-hour window and a weekly budget, and either one can stop a session. A verified Ultra subscriber works against the weekly budget only, so a large share of the week's usage can be spent in a single day. The trade-off runs in the other direction as well: an Ultra user who spends heavily early in the week reaches the weekly limit sooner and then waits for the weekly reset, which happens at a fixed time assigned to each account and shown under Settings > Usage.
··········
HOW THE ULTRA 5-HOUR EXEMPTION AND ACCOUNT VERIFICATION WORK.
The removal of the five-hour limit applies only to verified Ultra accounts. Anthropic verifies Ultra accounts automatically on the basis of signals such as account history and payment history, and there is no form or document upload. Accounts that are not yet verified carry a temporary five-hour limit, which Anthropic removes automatically once verification completes. The Ultra help article adds that confirming a phone number at checkout is a separate step and does not remove the temporary limit.
The check is visible in the product. Under Settings > Usage, a verified Ultra account no longer shows the "Current session" bar that tracks the five-hour window; if the bar is still present, the temporary limit is still in force. Anthropic has not published how long verification takes for a new subscriber, and the announcement on October 30 said the change was starting with most existing subscribers, which means some accounts that upgraded recently may still see the bar.
A related difference concerns Claude Code. The Claude Code help article describes a small wrap-up allowance that lets a response already in progress finish when the five-hour limit is reached, and on Max that allowance can apply each time the five-hour limit is hit, within the weekly limit. On a verified Ultra account the five-hour limit does not apply, so the interruption the wrap-up allowance is designed to soften does not occur.
··········
FABLE 5 AND FABLE 5.1 ACCESS ON MAX AND ULTRA.
Fable 5 and Fable 5.1 are Anthropic's most capable models, and their treatment is one of the main reasons people move from Pro to Max or Ultra. On Pro, neither Fable model is covered by plan limits and both run on pay-as-you-go usage credits from the first message. On Max and Ultra, both Fable models are included and can use up to 50% of the weekly usage limit at no extra cost.
........
Plan | Fable 5 and Fable 5.1 within plan limits | Fable cap | Fable cap in Pro-weeks |
Pro | No, usage credits from the start | Not applicable | 0 |
Max | Yes | Up to 50% of weekly limit | 2.5 |
Ultra | Yes | Up to 50% of weekly limit | 5 |
Team, premium seat | Yes | Up to 50% of weekly limit | Not expressed in Pro units |
Team, standard seat | No, usage credits | Not applicable | Not applicable |
........
The 50% figure is a cap on Fable use within the existing weekly limit; it does not add capacity. Fable usage counts against the same weekly budget as other models and, according to the Help Center, depletes it faster than other models. Because the cap is a share of each plan's weekly limit, Ultra's Fable budget is twice Max's in absolute terms. Once the Fable cap or the weekly limit is reached, either plan can continue on usage credits or switch to another model. At standard API rates, Fable 5.1 costs $10 per million input tokens and $50 per million output tokens, against $4 and $20 for Opus 5.5, which is the rate usage credits are billed at once the plan allowance runs out.
··········
MODELS AND FEATURES INCLUDED IN CLAUDE MAX AND CLAUDE ULTRA.
The feature lists of the two plans are almost identical. Anthropic's pricing page describes Ultra as everything in Max plus the Ultra-specific items: no five-hour limit for verified accounts, 10x Pro weekly usage and $200 in monthly API credits. Both help articles list priority access to new models, features and products, Claude Code under the same subscription, and longer multi-step tasks such as reports, spreadsheets and presentations that continue in the background.
........
Feature | Claude Max | Claude Ultra |
Opus 5.5, Sonnet 5.5, Haiku 5.5 | Included | Included |
Fable 5 and Fable 5.1 | Included up to 50% of weekly limit | Included up to 50% of weekly limit |
Claude Code (terminal, IDE, desktop, web) | Included, shared limits | Included, shared limits |
Background tasks in the unified Claude app | Included | Included |
Claude Duet research preview (US, mobile) | Included, one limit reset | Included, one limit reset |
Personal Finance beta (US) | Included | Included |
Priority access at high-traffic times | Included | Included |
Early access to advanced features | Included | Included |
Monthly Claude API credits | $100 | $200 |
........
The Claude Duet details show how closely Anthropic treats the two plans. Duet usage counts against the plan's five-hour and weekly limits at the same rate as chat, and during the research preview Max and Ultra each receive one limit reset, against three for Pro; unused resets expire on December 15, 2026. Duet requires a US account and is distributed through a waitlist. Claude Code's model configuration page lists Opus 5.5 as the default model for Pro, Max, Team and Enterprise, with a medium effort default set per model rather than per plan; the page does not yet mention Ultra by name.
··········
MONTHLY API CREDITS: $100 ON MAX AND $200 ON ULTRA.
The monthly API credits are the one benefit whose dollar value is fixed. Max includes $100 a month and Ultra $200 a month in credits for the Claude Platform, equal in both cases to the subscription price. The credits are claimed by linking a Claude Console organization under Settings > Billing, and they require the subscriber to have been on an eligible plan for seven days.
........
API credit rule | Claude Max | Claude Ultra |
Monthly amount | $100 | $200 |
Usable for | Messages API, Message Batches API, Console Playground, Managed Agents, Agent SDK | Same |
Not usable for | Interactive Claude Code, Claude Code GitHub Action, usage credits in Claude apps, Bedrock, Vertex AI, Microsoft Foundry | Same |
Rollover | None, expire at the end of each billing cycle | None, expire at the end of each billing cycle |
Upgrade from Max to Ultra | Not applicable | Prorated credits issued immediately, then $200 per cycle |
Linked organizations | One per plan | One per plan |
........
The restrictions limit who benefits. The credits pay for the subscriber's own applications and agents built on the Claude API, and they cannot extend the allowance used in Claude, Claude Code or Cowork. A Claude Code user who never calls the API directly gets no value from them, while a developer who runs an API-based agent, a batch job or the Agent SDK with an API key from the linked organization can offset the full subscription price if the credits are used each month. Because unused credits expire at the end of each cycle, the $100 difference between the plans only has value for a subscriber who spends at least that much on the API every month.
··········
WHAT CLAUDE MAX AND CLAUDE ULTRA COST PER UNIT OF USAGE.
Anthropic's multiples make a like-for-like comparison possible within its own plan ladder. The table below divides each plan's monthly price by its weekly allowance expressed in Pro units. The calculation was run in Python with the prices listed on November 7, 2026.
........
Plan | Monthly price | Weekly allowance (Pro = 1) | Price per Pro-week of capacity | 12-month cost | API credits per month |
Pro, monthly | $20 | 1 | $20.00 | $240 | None |
Pro, annual | $16.67 ($200 up front) | 1 | $16.67 | $200 | None |
Max | $100 | 5 | $20.00 | $1,200 | $100 |
Ultra | $200 | 10 | $20.00 | $2,400 | $200 |
........
The result is a flat rate across the paid ladder: the monthly Pro plan, Max and Ultra all charge $20 a month for each Pro-week of capacity, and only annual Pro is cheaper per unit. Ultra therefore carries no volume discount over Max. The premium a subscriber pays for Ultra buys the larger allowance, the absence of a five-hour window and the extra $100 of API credits, at the same unit rate as Max. Anthropic does not express Pro's allowance in tokens or messages, so these units measure relative capacity and cannot be converted into an absolute volume of work.
··········
WHAT $100 OF USAGE CREDITS BUYS COMPARED WITH UPGRADING TO ULTRA.
A Max subscriber who regularly runs out has two ways to spend the extra $100: upgrade to Ultra or buy usage credits, which are billed at standard API rates after a limit is reached. Usage bundles cut that rate: a $50 bundle costs $45, a $250 bundle costs $200 and a $1,000 bundle costs $700, with individual subscribers capped at $2,000 of discounted bundles a month.
To make the credits concrete, the table assumes a single long Claude Code session that processes 3 million input tokens, of which 80% are cache hits and 20% are written to the five-minute cache, and produces 150,000 output tokens. The token figures are an illustrative assumption, not an Anthropic measurement; the per-token prices are Anthropic's published API rates. The arithmetic was run in Python.
........
Model at API rates | Cost of one assumed session | Sessions covered by $100 of credits | Sessions covered by $200 of credits |
Sonnet 5.5 ($2 input, $10 output) | $3.24 | 30.9 | 61.7 |
Opus 5.5 ($4 input, $20 output) | $6.48 | 15.4 | 30.9 |
Fable 5.1 ($10 input, $50 output) | $15.60 | 6.4 | 12.8 |
........
The comparison favors Ultra for steady heavy use. $100 of usage credits at standard rates covers about 15 such Opus 5.5 sessions, and $100 spent on $50 bundles buys about $111 of credits at the 10% discount, while the same $100 spent on the upgrade doubles the weekly plan allowance and adds $100 of API credits. Anthropic does not publish what 5x Pro per week represents in tokens, so the exact break-even point cannot be calculated. The direction is clear from Anthropic's own pricing: a Max subscriber who already spends more than $100 a month on usage credits pays less on Ultra if the doubled weekly allowance absorbs that overflow, and a subscriber who exceeds Max only occasionally pays less by buying credits when needed.
··········
CLAUDE CODE ON MAX AND ULTRA.
Claude Code is included in both plans and shares the same limits as the Claude apps. Pro and Max subscribers sign in to Claude Code in the terminal, in supported IDEs such as VS Code, Cursor and JetBrains IDEs, and in the Code tab of Claude Desktop. An ANTHROPIC_API_KEY environment variable overrides the subscription and moves billing to the API, and the Help Center advises running /login to switch back to plan billing.
The five-hour window is where the plans diverge for developers. On Max, a long agentic session that reads a large repository, runs tests and edits many files can hit the five-hour limit while weekly capacity remains, and the session then waits for the window to reset or continues on usage credits. On a verified Ultra account, the same session runs until the weekly limit is reached. Both plans can run Fable 5.1 in Claude Code with the /model fable command (Claude Code 2.1.255 or later), within the 50% Fable cap.
Anthropic's Claude Code documentation has not fully caught up with the rename. The help article on using Claude Code with a subscription still tells Max 5x users to "upgrade to Max 20x" when they hit limits, and the model configuration page does not list Ultra among the plans. The pricing page and the Ultra help article are the current sources for the plan terms, and both describe the plans as set out in this comparison. Our guide to Claude plans and usage limits covers how Claude Code limits work across every tier, including Free, Pro, Team and Enterprise.
··········
CLAUDE MAX AND ULTRA COMPARED WITH TEAM PREMIUM SEATS.
Individual subscribers who work in a small group sometimes weigh Max against a Team premium seat. A Team premium seat costs $100 a month per seat on annual billing or $125 on monthly billing, with a minimum of two seats, and offers 5x the usage of a Team standard seat. Premium seats include Fable on the same 50% terms as Max and Ultra and carry $100 a month in pooled API credits per seat, with the team pool capped at $500 a month.
........
Item | Claude Max | Claude Ultra | Team premium seat |
Price | $100 a month | $200 a month | $100 a month annual, $125 monthly, minimum 2 seats |
Usage reference | 5x Pro | 10x Pro weekly | 5x Team standard seat |
Five-hour limit | Yes | None when verified | Yes |
Fable | Up to 50% of weekly limit | Up to 50% of weekly limit | Up to 50% of weekly limit |
API credits | $100 a month | $200 a month | $100 per seat, pooled, cap $500 |
Claude Duet | Included | Included | Not available yet |
........
The Team plan adds administrative features such as single sign-on and central billing, which Max and Ultra do not offer, and it is the only one of the three that can be billed annually. Team does not have an equivalent of Ultra's exemption from the five-hour limit, and Claude Duet is not yet available on Team or Enterprise.
··········
WHAT ANTHROPIC HAS NOT PUBLISHED ABOUT MAX AND ULTRA LIMITS.
Several numbers that would sharpen the choice are not public. Anthropic does not state Pro's allowance in messages, tokens or hours, so the 5x and 10x multiples are relative measures whose real size depends on the model, the effort level, conversation length and file size. The ratio between Pro's five-hour allowance and its weekly allowance is not published either, which means the point at which Max's five-hour cap binds before its weekly cap cannot be computed from official figures.
Anthropic has not published how long Ultra verification takes, which signals beyond account and payment history it uses, or when the October 30 rollout reached every existing subscriber. The Help Center also reserves the right to add weekly or monthly caps or model and feature restrictions on both plans without fixed notice, and it does not guarantee uninterrupted service. No independent organization measures consumer plan allowances, so the comparison rests on Anthropic's own pages and on the published API rates used in the cost example above. Readers comparing Claude's top tier with other vendors' entry plans can find the cross-vendor view in our comparison of ChatGPT Plus, Claude Pro and Google AI Pro.
··········
CLAUDE MAX OR CLAUDE ULTRA: WHICH PLAN FITS WHICH USER.
The choice comes down to two questions: whether the five-hour window interrupts real work, and whether usage regularly overflows 5x Pro per week. The table maps common usage patterns to the plan that Anthropic's published terms favor.
........
Usage pattern | Recommended plan | Reason |
Daily chat, writing and analysis that outgrows Pro | Max | 5x Pro at the same unit price, rarely limited by the 5-hour window |
Claude Code in short, frequent sessions | Max | Weekly budget is the binding limit, Ultra's exemption adds little |
Long Claude Code or agent runs that stop at the 5-hour cap | Ultra | No 5-hour limit once verified, 10x Pro weekly |
Heavy Fable 5.1 use | Ultra | Fable cap of 5 Pro-weeks against 2.5 on Max |
Max user spending over $100 a month on usage credits | Ultra | Same $100 buys double the weekly allowance plus $100 more API credits |
Max user exceeding limits a few times a month | Max with usage credits or bundles | Credits cost less than the $100 upgrade for occasional overflow |
Developer who builds on the Claude API | Max or Ultra by API spend | Credits equal the plan price, but only if used every month |
Small group needing SSO or annual billing | Team premium seats | Admin features and annual pricing, no Ultra-style 5-hour exemption |
........
For most subscribers who already find Pro too tight, Max is the sensible starting point, since it costs the same per unit of usage as Ultra and leaves the option to upgrade with prorated billing. Ultra earns its $200 price for verified users whose long Claude Code sessions run into the five-hour cap, for those who lean on Fable 5.1 for a large share of their work, and for Max subscribers whose monthly usage-credit spend already exceeds $100.
··········
FOLLOW US FOR MORE.
·····
DATA STUDIOS
·····
datastudios.org




