Claude Max Pricing Guide · Verified October 27, 2026
Max 5x costs $100 a month and Max 20x costs $200, billed monthly only. The 5x and 20x apply to each five-hour session, not to your week. Since October 7, both plans also include Claude API credits. Every figure is sourced, and the full list is at the end.
Max gives you more Claude. Suprmind gives you a second opinion on it: Claude answers in one thread with GPT, Gemini and Grok, each model reads what the others said, and a wrong answer can be challenged before you act on it.
Plans from $19 a month. Perplexity joins as the fifth model from Pro at $45.
In this update: API credits on Max, Claude Code's weekly limits after the September 14 change, the Opus 5.5 limit increase, ChatGPT Pro's new $500 tier and the Kahn v. Anthropic hearing set for November 6.
Claude Max is Anthropic's plan for heavy individual use. It runs the same Claude models as Pro and sells you more usage before a limit stops you. These are the questions people search for most, each checked against Anthropic's own pages on October 27, 2026.
$100 a month for Max 5x and $200 for Max 20x, plus tax. Monthly billing only. Pro is $20, or $17 a month billed yearly.
Five or twenty times Pro's usage per five-hour session. A weekly limit sits on top, and it grows far less than the session limit.
No. Max is monthly only, and Anthropic offers no standard discounted pricing on Pro or Max.
Since October 7, Max 5x includes $100 and Max 20x $200 of Claude API credits a month, for the API and the Agent SDK. Not for interactive Claude Code. They expire each cycle.
Yes, one pool. But most of the limit increases in 2026 applied only to Claude Code.
Yes, up to 50% of your weekly limit. On Pro, only with paid usage credits. Mythos 5.1 is not part of Pro or Max.
You wait for the reset, or keep going on usage credits billed at API rates. Prepaid bundles take up to 30% off.
If your Claude use would cost more than $100 a month at API rates, yes. Published reports put a heavy month of Claude Code at $1,200 to $1,600 at API rates.
Max lets one model answer more questions. It can't tell you when that model is wrong. In Suprmind, Claude answers in the same thread as GPT, Gemini and Grok. Each model reads what the others said, so a weak answer can be challenged right there, before you act on it.
See Where the Models DisagreePlans from $19 a month. Compare every Suprmind plan.
Anthropic sells three individual plans and two kinds of Team seat. Max is the only individual plan sold monthly only, and there is no plan above Max 20x. Prices are Anthropic's web prices before tax, from its pricing page and the Max plan help article.
Team seats come in two sizes, and a Premium seat is the closest thing to Max inside a company account: five times the usage of a Standard seat. Enterprise is a seat fee plus usage.
| Plan | Price | Seats | Usage | Monthly API credits |
|---|---|---|---|---|
| Team Standard seat | $25 a seat a month, or $20 billed yearly | 2 to 150 | Team baseline | $20 a seat, pooled up to $500 |
| Team Premium seat | $125 a seat a month, or $100 billed yearly | 2 to 150 | 5x a Standard seat | $100 a seat, pooled up to $500 |
| Enterprise | $20 a seat a month billed yearly, plus usage | 20-seat minimum | Billed by model and task | None |
Max runs the same models as Pro. On top of the larger limits, Anthropic lists higher output limits for all tasks, early access to advanced Claude features and priority access at high traffic times. Max can also spend up to half its weekly limit on Fable models, which Pro can only use on paid usage credits, and it now includes monthly API credits. Context windows depend on the model, not the plan: the newest models support up to 1M tokens, others 500K or 200K, per Anthropic's limits article.
Anthropic's pricing FAQ defines the multiplier in one line: "5x or 20x the usage of Pro per 5-hour session." Above the session sits a weekly limit across all models, and it does not scale by 5 or 20. Anthropic published weekly figures once, in subscriber emails when weekly limits arrived in July 2025. Since then, the other figures come from a lawsuit's allegations and from independent tests. Here they all are on one scale.
Anthropic's weekly figures are ratios of the hour ranges in its July 2025 subscriber emails (Sonnet 4 and Opus 4), quoted in the Kahn complaint and in TechCrunch. Anthropic has not published weekly ranges for current models. Independent tests are single measurements, not averages.
Switch to the second view. Max 20x costs twice as much as Max 5x and gives four times as much per session. Per week, Anthropic's own 2025 estimates gave it 1.7 times the Sonnet hours and 1.1 to 1.6 times the Opus hours. A test by Botfarm in early 2026 measured about twice the weekly usage, and its author summed it up: "The $200 plan gives you 4x the burst capacity but only ~2x the sustained weekly budget," per Botfarm. So the real advantage of Max 20x is depth: one long session that keeps going.
The complaint says the Max 20x checkout carries a "Save 50%" label. That saving holds per session: Max 20x costs $10 for each Pro's worth of usage against $20 on Max 5x. Measured by the July 2025 weekly hours, each Sonnet hour cost about 17% more on Max 20x than on Max 5x, and about 43% more on Max 5x than on Pro at its monthly price.
| Where you get stopped | What it means | What to buy |
|---|---|---|
| The five-hour window, mid-task | One long session uses it up. You need depth. | Max 20x. Its 4x session gain is real. |
| The weekly limit, before the week ends | Total volume is the problem, not session depth. | Max 20x adds about 2x a week at best. Usage credits for heavy weeks often cost less. |
| Neither, or only now and then | Pro's limits are close to enough. | Pro billed yearly, plus usage credits when a week runs hot. |
Watch Settings, Usage for two normal weeks before you pick. It shows the current session, the weekly limit and your next reset time.
Two Max 5x seats cost the same $200. By Anthropic's July 2025 estimates they held 280 to 560 Sonnet hours a week between them, against 240 to 480 on one Max 20x. "This means 2 $100 plans provides more usage than 1 $200 plan. Very misleading," one Hacker News commenter wrote in August 2026. One Max 20x still wins for a single long session, with four times the usage per five-hour window of either 5x seat.
The catch is who uses the seats. Two seats for two people is ordinary. One person running two accounts to get around limits is a risk to both accounts: Anthropic's Claude Code docs say advertised limits "assume ordinary, individual usage", and its Usage Policy, in the version effective November 12, 2026, bars creating multiple accounts to circumvent product guardrails. The terms do not address one person paying for two plans directly.
Kahn v. Anthropic PBC was filed on June 14, 2026, in the U.S. District Court for the Northern District of California (No. 3:26-cv-05763). The complaint alleges that Max 5x delivers about 3.5 times Pro's usage and Max 20x six to eight times, that "Defendant delivers only 1.1 to 1.6 times the usage of Max 5x with the Max 20x subscription plan," and that the per-hour price of Max 20x is higher than on Max 5x or Pro.
The 3.5x, 6x and 1.1 to 1.6x figures are ratios of the hour ranges in Anthropic's own July 2025 subscriber emails. Anthropic's answer, in its August motion to dismiss: the 5x and 20x are promised per session, and "per-session usage and weekly usage limits are, on the face of the statement, two different limits."
In August the court found a second suit, Smith v. Anthropic, related to Kahn, and in September the cases were consolidated. A consolidated complaint followed on September 8. Anthropic has moved to dismiss it, and the hearing is set for November 6, 2026, before Magistrate Judge Alex G. Tse, per the court docket. These are allegations. No court has ruled on them, and no class has been certified. A separate group of cases, In re Claude Service Degradation Litigation, covers Pro and Max subscribers who used Claude between March 4 and May 6, 2026, and runs on its own track.
Every paid Claude plan runs two limits at once: a session limit that resets every five hours, and a weekly limit that resets once a week. The 5x and 20x apply to the first. Most complaints about Max come from the second.
This is where the multiplier lives: five or twenty times Pro's usage in each five-hour session. When you reach it, you wait for the reset or keep working on usage credits, per Anthropic's limits article.
Max plans also have a weekly limit that applies across all models. It resets at a fixed time each week, assigned to your account, and Settings, Usage shows when. Anthropic announced weekly limits in July 2025, effective August 28, and said they would affect less than 5% of subscribers based on usage at the time.
Max plans can spend up to 50% of the weekly limit on Fable models at no extra cost. That is a cap, not an extra pool: Fable draws from the same weekly limit, and Anthropic notes that Fable models "draw from your plan's regular weekly usage limits and use them faster than other Claude models." On Pro, Fable sits outside the plan's limits and runs only on usage credits. Mythos 5.1 is not part of Pro or Max. Anthropic offers it through trusted access programs. Details: Claude Fable models on your plan.
Anthropic names the length and complexity of your conversations, the features you use, the model you chat with and the effort level you select. Tools and connectors are "token-intensive," and long conversations that trigger automatic context management use more of the limit. There is no fixed message count on any plan. In March 2026, The AI Economy reported that limits run down faster at weekday peak hours. On May 6, Anthropic removed that peak-hours reduction for Claude Code on Pro and Max.
With usage credits turned on, your work continues on them instead of stopping. They are billed at standard API rates, per Anthropic's usage credits article. Prepaid usage bundles cut that price by up to 30%: a $1,000 bundle costs $700, and you can buy up to $2,000 worth of discounted bundles a month. For a week of heavy work once a month, a bundle usually costs less than moving up a plan.
Most of the limit increases Anthropic announced in 2026 applied only to its agent products, Claude Code and Cowork.
Claude Code shares the same usage limits as the rest of your plan. In Anthropic's words on its pricing page, "your work in the terminal and your chats draw from one pool." Max was not built for this. On April 21, 2026, the day Anthropic briefly tested dropping Claude Code from Pro for some new signups, its Head of Growth, Amol Avasare, wrote:
"When we launched Max a year ago, it didn't include Claude Code, Cowork didn't exist, and agents that run for hours weren't a thing. Max was designed for heavy chat usage, that's it."
Amol Avasare, Anthropic, April 21, 2026
The pages were reverted the same day, and Claude Code is still part of Pro, per Where's Your Ed At and Simon Willison. The gap Avasare described explains most of 2026: agents that run for hours burn through limits sized for chat, and Anthropic has spent the year adjusting Claude Code's limits instead of the price.
Anthropic publishes its own figure from enterprise deployments: "the average cost is around $13 per developer per active day and $150-250 per developer per month, with costs remaining below $30 per active day for 90% of users," per its Claude Code cost docs. At those rates a typical month of Claude Code costs more than Max 5x and about what Max 20x costs. Heavy users run far past both.
ccusage is a community tool, not Anthropic's. It reads Claude Code's local logs and prices them at API rates. Run npx ccusage@latest monthly in a terminal. The total is what your month would have cost on the API, and it is the number the calculator below needs.
Anthropic's Help Center says you can still use the Claude Agent SDK, claude -p and third-party apps with your subscription limits. What it does not allow, per its Claude Code legal page, is a third-party developer routing its own users' requests through their consumer plan logins. In April 2026 Anthropic cut subscribers off from OpenClaw, a popular open-source agent, in what The Next Web called a cost crackdown. The new API credits are a separate pot: they work in the Agent SDK and in third-party harnesses that accept a Claude API key, but they "don't cover interactive Claude Code sessions," per Anthropic's Help Center.
From the founder
My last month on the API cost $3,600. Today I pay $400.
Before I moved to Max 20x, I was running Claude Code in Cursor on my own API key. The last month of that setup cost me $3,600 in one calendar month, for API tokens alone.
Then I realized that Max covers Claude Code as well, and that was the most beautiful thing I ever learned. Only half joking.
By the time I started building Suprmind, I was already on Max 20x. At the peak I paid for three seats, $600 a month. Today it's two, $400 a month. That one API month would cover nine months of what I pay now.
I like Max. For building, it's the best deal in AI right now. But paying for more of one model taught me what more of one model doesn't buy: a second opinion. In my experience, Claude sounds just as sure when it's wrong as when it's right. When a decision actually matters, I want other models reading the answer and pushing back before I act on it.
That's why we built Suprmind. Max is how I build. Suprmind is how I check what I'm about to build.
Max makes Claude cheap to run. It can't tell you when Claude is wrong. In Suprmind, Claude answers in one thread with GPT, Gemini and Grok. Each model reads what the others said, so a weak answer can be challenged before you ship it.
Cross-Check Claude With Other AIsPlans from $19 a month. Compare every Suprmind plan.
Plans are sold by usage, not dollars, so the fair test is what your usage would cost at API rates. Enter that number from ccusage or the Claude Console, or estimate it from tokens. The calculator compares it with each plan and counts the new API credits if you also pay for the API.
API prices from Anthropic's pricing page, checked October 27, 2026. Token mode prices new input at the 5-minute cache-write rate and cached input at the cache-hit rate, which is how Claude Code bills on the API.
| Model | Input | Output | Cache read | Note |
|---|---|---|---|---|
| Claude Opus 5.5 | $4 | $20 | $0.20 | |
| Claude Sonnet 5.5 | $2 | $10 | $0.10 | |
| Claude Fable 5.1 | $10 | $50 | $0.25 | On Max: up to 50% of the weekly limit |
| Claude Haiku 5.5 | $0.10 | $0.50 | $0.01 | Prompts up to 100K tokens. Above that: $0.50 in, $2.50 out |
For coding agents, most of the bill is cached input, and cache reads got cheap: 5% of the input price on Opus 5.5 and Sonnet 5.5, and 2.5% on Fable 5.1. The Batch API halves input and output prices for work that can wait. Haiku 5.5 costs five times more on prompts over 100,000 tokens, and US-only inference adds 10% on current models.
Since October 7, 2026, Max 5x includes $100 and Max 20x $200 of Claude API credits each month. Team plans get $20 per Standard seat and $100 per Premium seat, pooled and capped at $500. Anthropic announced it with Haiku 5.5, and the rules are in its platform docs:
What this changes: if you already pay for the API, Max now covers part of that bill. A developer with $150 of monthly API usage on Max 5x pays $100 for the plan and $50 for the API, instead of $250 for both. With no API usage, the credits are worth nothing to you.
Log scale. Weekly test results are multiplied by 4.33 to show a month. Every report predates at least one limit change, so read them as orders of magnitude.
Two kinds of numbers get quoted, and they answer different questions. Some people price their own month. Simon Willison on Max 5x came to $1,199.79 of Claude Code usage at API rates over 30 days in May 2026, and Paweł Huryn on Max 20x to $1,588 in a month of Claude Code. Others test what a plan holds when it is maxed out. Botfarm measured about $523 a week on Max 5x and about $1,100 on Max 20x in early 2026, almost all Opus 4.6, before the May limit changes. SemiAnalysis ran long coding tasks to the weekly limit and reported that Max 20x delivers about $8,000 of API-equivalent tokens a month, per Pasquale Pillitteri's summary.
Either way the gap is large. The two personal months came to 8 and 12 times the plan price at API rates, and the maxed-out tests to 23 to 40 times. That gap is why many people who code with Claude every day pick Max.
For people who run Claude Code most days, the numbers say yes, by a wide margin. "I pay $200/month for Claude Max and I consider it well worth the money," Simon Willison wrote in April 2026, and in October SemiAnalysis found that Anthropic's subscriptions offered about five times the API-equivalent value of OpenAI's for mid-tier models, with similar limits on the top models. The complaints are real too: a Hacker News thread titled "Pro Max 5x quota exhausted in 1.5 hours despite moderate usage" drew 759 points in April. Your own pattern decides which camp you land in.
That last point is the one Suprmind is built for. More Claude doesn't catch Claude's mistakes. On the AA-Omniscience benchmark, Claude Opus 5.5 gives a wrong answer instead of declining 58.6% of the time when it can't answer correctly, and Fable 5.1 does it 72.6% of the time, per our AI hallucination rates page. That is not a Claude-only problem: Opus 5.5 still has the highest AA-Omniscience Index in that table, and GPT-6 Astra's rate is 51.3%. A second model reading the answer can catch what the first one won't flag. Where Suprmind fits explains how.
Every major lab now sells a tier at $100 to $200 a month, and Google's mirrors Max's structure almost exactly. This is what each buys as of October 27, 2026.
| Plan | Price a month | What the price buys | Built for |
|---|---|---|---|
| Claude Max 5x | $100 | 5x Pro per five-hour session, plus a weekly limit. $100 of API credits a month. | Claude Code and long Claude sessions |
| Claude Max 20x | $200 | 20x Pro per five-hour session, plus a weekly limit. $200 of API credits a month. | All-day Claude Code and long agent runs |
| ChatGPT Pro, $100 tier | $100 | 5x ChatGPT Plus usage. No five-hour limit on Pro plans, per OpenAI's Codex lead. | Codex and heavy ChatGPT use |
| ChatGPT Pro, $200 tier | $200 | 10x Plus for anyone subscribing after September 29, 2026, and for earlier subscribers from October 30. It was 20x. | Heavy Codex use |
| ChatGPT Pro, $500 tier | $500 | 25x Plus, as reported by Windows Report, plus Ultrafast versions of OpenAI's top models. | The heaviest OpenAI users |
| Google AI Ultra | $99.99 or $199.99 | 5x or 20x the usage of AI Pro. Resets every five hours, up to a weekly limit. | Gemini, Jules and Antigravity |
| X Subscription Ultra | $200 | 50x the $8 Plus tier, in one weekly pool shared by Grok and the Cursor agent. | Grok and Cursor users |
| Perplexity Max | $200, or $2,000 a year | Models from OpenAI, Anthropic and others. The highest limits on its browser agent. | Research and browsing agents |
| Suprmind Frontier | $95 | GPT, Claude, Gemini, Grok and Perplexity in one thread, each reading the others. Debate, Red Team and decision briefs. | Decisions checked by five models |
| Suprmind Power | $195 | Everything in Frontier, the most usage of any self-serve Suprmind plan, and your own API keys. | Heavy multi-model work |
One contrast matters for coders. On August 31, 2026, the day a Hacker News thread pointed out that Max 20x's multiplier stops at the session, OpenAI's Codex lead Tibo Sottiaux posted: "For clarity, while both are called 20X, in Codex they apply specifically to weekly usage limits. And we also don't have 5h limits for both Pro plans. The Pro 20X is quite precisely 20X the usage of the Plus subscription, so it does exactly what it says on the tin." Four weeks later OpenAI cut that tier to 10x.
Suprmind's Frontier ($95) and Power ($195) plans sit at the same prices with a different job. They don't sell more of one model. They put GPT, Claude, Gemini, Grok and Perplexity in one thread, reading and checking each other.
Search for cheap Claude Max and you'll find shared seats, resold logins and API relays priced far below Anthropic's list price. ChinaTalk documented how it works in May 2026: one $200 Max plan carved up among several users through tokens-per-hour quotas, with resale at 70 to 90% below official prices. Here is what that money buys.
Anthropic says it doesn't offer standard discounted pricing on Pro or Max. It occasionally runs limited-time promotions, so check any discount code against Anthropic's own site before you pay for it.
The price has not moved since April 2025. Almost everything else has. Filter by type and follow each entry to its source.
API credits added to Max
Max 5x gets $100 and Max 20x $200 of Claude API credits a month. Team seats get $20 or $100, pooled up to $500. Anthropic
One-time cloud session credit
Existing subscribers get a one-time credit to try cloud sessions: $100 on Pro and $250 on Max. @ClaudeDevs
Opus 5.5 raises five-hour limits
Opus 5.5 launches with higher five-hour limits on Pro, Max, Team and seat-based Enterprise plans, 20% higher in Claude Code. Subscribers also get a one-time limit reset to use whenever they choose. @ClaudeDevs
Claude Code weekly promotion ends
Weekly limits in Claude Code are now 25% higher than they were before the promotion. Claude Help Center
Cases consolidated
The court approves a stipulation consolidating the Max cases. A consolidated complaint follows on September 8, and Anthropic files a new motion to dismiss on September 22. ECF No. 35
Fable 5.1 released
Max can use it within the 50% Fable cap. Mythos 5.1 stays limited to trusted access programs. Anthropic
Weekly limit change announced
Anthropic says Claude Code weekly limits will settle 25% above the pre-promotion level from September 14, which it calls a 17% reduction from the promotion. @ClaudeDevs
Smith v. Anthropic related to Kahn
The court finds a second suit, Smith v. Anthropic (No. 4:26-cv-07789), related to Kahn. On August 19, Anthropic moves to dismiss, with a hearing set for November 6. ECF No. 23
Fable capped at 50% of the weekly limit
After a launch promotion ends on July 19, Max plans can spend up to half the weekly limit on Fable models. Pro needs usage credits for Fable. Claude Help Center
Fable 5 access restored
The export controls are lifted on June 30, and Fable 5 returns on July 1 in Claude, Claude Code and on the Claude Platform. Anthropic
Kahn v. Anthropic filed
A proposed class action in the Northern District of California alleges Max delivers far less than the advertised 5x and 20x. the complaint
Fable 5 access suspended
Anthropic suspends access to Fable 5 and Mythos 5 after a U.S. government export control directive citing national security. Other Claude models are not affected. Anthropic
Cowork five-hour limits doubled
A promotion doubles Claude Cowork's five-hour limit on Pro, Max and Team through July 5. Chat, Claude Code and weekly limits do not change. Claude Help Center
Claude Code weekly limits up 50%
A promotion raises Claude Code's weekly limits by 50%, first through July 13 and in the end through September 13. Five-hour limits are not affected. @ClaudeDevs
Claude Code five-hour limits doubled
Doubled on Pro, Max, Team and seat-based Enterprise, and the peak-hours reduction for Claude Code ends on Pro and Max. The same post announces a compute deal for SpaceX's Colossus 1 data center. Anthropic
Claude Code briefly dropped from Pro
Anthropic's pricing page and docs remove Claude Code from Pro in a test on about 2% of new signups, then revert the pages the same day. Claude Code stays in Pro. Where's Your Ed At
Third-party agents cut off
Anthropic cuts subscriber logins off from OpenClaw, a popular open-source agent. Its Claude Code legal page bars third-party developers from routing their users' requests through consumer plan logins. The Next Web
Peak-hour limits reported
The AI Economy reports that limits run down faster at weekday peak hours. The AI Economy
Opus 4.5 removes Opus-only caps
With Opus 4.5, Anthropic removes Opus-specific caps and raises overall usage limits for Max and Team Premium. Anthropic
Weekly limits take effect
Pro and Max now have a weekly limit across all models on top of the five-hour session limit. TechCrunch
Weekly limits announced
Anthropic says weekly limits start August 28 and will affect less than 5% of subscribers. Its emails estimate 140 to 280 hours of Sonnet 4 and 15 to 35 of Opus 4 a week on Max 5x, and 240 to 480 and 24 to 40 on Max 20x. TechCrunch
Claude Max launches
Anthropic adds Max at $100 a month for 5x Pro's rate limits and $200 for 20x. TechCrunch
We pay for Claude Max ourselves, and we would buy it again. It is the right tool when you need volume from one model: code, long build sessions, drafts. Suprmind does a different job: the calls you want checked before you act on them.
Why a second model matters even with the best Claude: in our Multi-Model AI Divergence Index, Claude was in the group other models corrected least often, with GPT and Perplexity. Another model still corrected 4.3% of Claude's answers when two of three AI judges agreed, and 8.9%, about 1 in 11, with a single judge. When two fact-check judges from different companies agreed, they sided with the correcting model 18 times out of 23.
One thing Claude does better on its own: it writes and edits a document right in the conversation. In Suprmind, documents come from the Master Document Generator, a separate step where you pick the template and the model that writes it.
Suprmind plans are $19 (Spark), $45 (Pro), $95 (Frontier) and $195 (Power) a month. Debate, Red Team and the decision tools start on Pro.
Max 5x costs $100 a month and Max 20x costs $200 a month, plus tax where it applies. Both are billed monthly only. In the iPhone app, the App Store lists them at $124.99 and $249.99, so buying on the web is cheaper.
No. Max is sold monthly only, and Anthropic says it doesn't offer standard discounted pricing on Pro or Max, though it occasionally runs limited-time promotions. Pro is the only individual plan with yearly billing: $17 a month, paid as $200 up front. Nonprofits can buy Team seats at $8 a user a month, and Anthropic's open-source program covers six months of Max 20x for maintainers it accepts.
Per five-hour session, Max 5x gives five times Pro's usage and Max 20x gives twenty times. Both have a weekly limit on top that grows much less. In Anthropic's July 2025 estimates, Max 20x got about 1.7 times the weekly Sonnet hours of Max 5x and 1.1 to 1.6 times its Opus hours. The models and features are the same. Max 20x includes $200 of API credits a month against $100 on Max 5x.
No. The 4x applies to each five-hour session. Per week, Anthropic's own July 2025 figures put Max 20x at 1.1 to 1.7 times Max 5x, and a test by Botfarm in early 2026 measured about 2.1 times. If you run into the weekly limit rather than the session limit, Max 20x helps less than its price suggests.
Yes, since October 7, 2026. Max 5x includes $100 and Max 20x includes $200 of Claude API credits each month. They cover the Claude API, Managed Agents, the Agent SDK and the playground, but not interactive Claude Code sessions or extra usage in the Claude apps. Unused credits expire at the end of each billing cycle, and new subscribers can claim them after 7 days.
Yes. Anthropic says Claude Code shares the same usage limits as the rest of your plan, so work in the terminal and your chats draw from one pool. One difference: most of the limit increases Anthropic announced in 2026 applied only to Claude Code.
You wait for the five-hour window or the weekly limit to reset, or you keep working on usage credits. Usage credits are billed at standard API rates, and prepaid bundles cut that price by up to 30%: a $1,000 bundle costs $700. Settings, Usage shows your current usage and your next reset time.
Fable, yes. Max plans can spend up to 50% of their weekly limit on Fable models at no extra cost, and that use counts toward the weekly limit. On Pro, Fable runs only on usage credits. Mythos 5.1 is not part of Pro or Max. Anthropic offers it through trusted access programs.
No. Max 20x at $200 a month is the largest individual Claude plan. The only $500 figure on Anthropic's pricing page is the cap on a Team's pooled API credits. OpenAI does sell a $500 ChatGPT Pro tier.
For heavy use, by a wide margin. Anthropic says Claude Code averages about $150 to $250 per developer per month across enterprise deployments, and published reports of a heavy month on Max run from about $1,200 to $1,600 at API prices. If your use would cost less than $100 a month at API rates, the API or Pro is cheaper.
Prices in the iPhone app are set through Apple's App Store, which lists Max 5x at $124.99 and Max 20x at $249.99 a month. On the web they cost $100 and $200. Plans bought through the app still get the monthly API credits, but only Apple can refund them.
Not within Anthropic's terms. The Consumer Terms say you may not share your login or make your account available to anyone else, and you may not resell the services. Shared and resold Max accounts break both rules, can be closed without notice, and pass your prompts through someone else's account.
Usually not. Anthropic treats payments as non-refundable, and you should cancel at least 24 hours before your renewal date to avoid the next charge. Buyers in the EEA and the UK get a prorated refund within 14 days of purchase. If you subscribed in the iPhone app, only Apple can issue a refund.
It depends on how you hit limits and who uses the seats. Two Max 5x seats cost the same $200 and, by Anthropic's July 2025 estimates, held more combined weekly hours. One Max 20x gives four times as much per five-hour session as either 5x seat. Two seats for two people is ordinary. One person running two accounts to get around limits risks both accounts.
Kahn v. Anthropic PBC, filed June 14, 2026 in the Northern District of California, alleges that Max delivers far less than its multipliers suggest: about 3.5 times Pro on Max 5x, six to eight times Pro on Max 20x, and only 1.1 to 1.6 times Max 5x on Max 20x. A related suit, Smith v. Anthropic, was consolidated with it in September, and a consolidated complaint followed on September 8. Anthropic argues that the 5x and 20x apply per session and that the weekly limit is a separate limit. It has moved to dismiss, and the hearing is set for November 6, 2026. These are allegations. No court has ruled on them.
They cost the same and are built differently. ChatGPT Pro's $100 and $200 tiers have no five-hour limit, and OpenAI sells usage as a multiple of Plus: 5x at $100, 10x at $200 since OpenAI cut that tier from 20x this fall, and a reported 25x at $500. Claude Max multiplies Pro per five-hour session and adds a weekly limit. In October 2026, SemiAnalysis found Anthropic's subscriptions gave about five times the API-equivalent value of OpenAI's for mid-tier models, with similar limits on the top models. For coding, the agent you prefer, Claude Code or Codex, usually decides.
Mostly no. Pro and Max run the same Claude models. Max adds usage, priority at busy times, higher output limits and early access to features. The exception is Fable: Max can spend up to half its weekly limit on Fable models, while Pro pays for Fable with usage credits.
Open Settings, Usage. It shows your current session, your weekly limit and the time of your next weekly reset. For Claude Code, the community tool ccusage reads your local logs and prices them at API rates, which is the number to compare with the plan price.
Every price, limit and date was checked first against Anthropic's own pages: the pricing page, the Help Center, the platform and Claude Code docs, Anthropic's announcements and the official @AnthropicAI and @ClaudeDevs posts. Court details come from the public docket. Competitor prices come from each company's own pricing or help pages, with press reports only where a company does not publish a figure. User reports are labeled as reports, with the basis of each. Where a source is older than a limit change, the page says so.
Every chart reads one data file, so a correction there updates the page and every embedded copy. Download the data (JSON). Verified October 27, 2026. We re-check it every month and whenever Anthropic changes a plan.
Keep Claude Max for the volume. When a call matters, put Claude in one thread with the other frontier models and see where they disagree before you act.
