Claude Pricing Explained (2026): What Pro, Max and Team Actually Cost

Claude Pro costs $20 a month, or $200 a year paid up front. Claude Max costs $100 a month for Max 5x and $200 a month for Max 20x, and it is sold monthly only. Those are the prices on Anthropic's own pages when I read every plan card and help article on 13 October 2026.

The rest of the ladder, as of October 2026: Free is $0. A Team seat costs $25 (Standard) or $125 (Premium) a month on monthly billing, for teams of 2 to 150. Enterprise is $20 a seat a month, billed annually, with usage charged on top at API rates.

Two details on the 13 October page catch people out. The "$17" on the Pro card is the annual price divided by twelve, so no plan bills $17 a month. And the pricing page never prints the Max 20x price; it says "From $100" and leaves the $200 to a help article.

Usage is where the real decision sits, and Anthropic publishes no message count for any plan. Below: every plan on each billing cycle it offers, what the usage multiples actually compare, and which plan I would pick for which kind of week. Still choosing an assistant in the first place? Our AI assistant for work comparison prices the wider field.

How I read the prices

I read claude.com/pricing six times on 12 October 2026 and seven more times on 13 October, through US and Indian connections, some of them via anthropic.com/pricing, which now redirects there. Every price shown matched across all thirteen loads, even though the page was testing its layout: a Developer tab on some loads, a "Claude Code included" badge on others.

There is no monthly or annual toggle. For what the cards leave out, I read eight Anthropic help and docs pages on 13 October, all listed in the sources.

Three caveats apply to the plan prices here. "Prices shown don't include applicable tax." Max's prices "are for web subscriptions only. Mobile pricing may vary depending on your app platform." And loads from India still showed US dollars, though the Pro help article promises "pricing in your local currency where supported", so I quote USD only.

Claude plans at a glance

Seven ways to pay, as of 13 October 2026. The last column uses Anthropic's own allowance wording.

Prices as displayed on claude.com/pricing on 13 October 2026; Individual-tab prices exclude tax. Max 20x's $200 comes from Anthropic's Max help article; the pricing page shows only "From $100". The Team multiples against Pro come from the Team help article. Quoted taglines are Anthropic's; the Max 20x and Team Premium fit lines are mine.

Claude Pro: $20 a month, or $200 a year up front

The Pro help article prices it at "$20 per month (US)" and notes that "Monthly pricing varies by region". As of October 2026, the annual option is $200 billed up front.

Against Free, the card and the comparison table add five things:

• More usage: at least 5x Free's allowance per five-hour session.
• Research and Claude Code, both marked "No" on Free.
• Opus. Sonnet and Haiku are on every plan, but Opus starts at Pro.
• Projects, where Free stops at "Up to 5".
• The card's extras: "Hand off and schedule tasks", Claude Dashboards (beta), Claude Science, and Claude in Chrome and Microsoft 365.

Pro leaves out three things. Its Fable cell reads "Usage credits", the pay-as-you-go balance explained further down, where both Max columns read "50% of weekly limits*". It gets no monthly API credits. And the help article is blunt: "The Pro plan does not include API usage through the Claude Console."

One line I cannot settle for you. The comparison table marks Pro "No" for "Priority access at high traffic times", while the Pro help article lists "Priority access to Claude during high-traffic periods" as a Pro benefit, and both said so on 13 October 2026. If priority matters, Max is where the card and the table agree.

Annual Pro saves $40, and refunds are the exception

A year of monthly Pro costs 12 x $20, which is $240. Annual costs $200, so the saving is $40 as of October 2026, exactly two monthly payments.

The catch is the refund wording: "Payments are generally non-refundable, except where our Consumer Terms of Service say otherwise or where local law requires it." Customers in the EEA and UK get a 14-day withdrawal refund in the app. Everywhere else there is an eligibility check through Get help, and no refund window is stated.

So pay monthly first. Buy the year when you would pay it without thinking.

Claude Max vs Pro: 5x or 20x, monthly only

Max comes in two sizes, and the pricing page shows only the first: "From $100 Per month", with "Choose 5x or 20x more usage than Pro*". The Max help article, read on 13 October 2026, finishes the thought: "Max 5x: $100 per month. Max 20x: $200 per month."

Monthly only. The help article says "The Max plan is currently available as a monthly subscription only", and the pricing page's billing row reads "Monthly" for both tiers. No Anthropic page I read on 13 October 2026 offers Max on annual billing.

Price, as of October 2026

$100 a month (Max 5x) or $200 a month (Max 20x), web subscriptions.

Usage

5x or 20x Pro's per-session allowance, plus a weekly limit across all models.

Fable

"50% of weekly limits*", plus a separate weekly limit for Fable.

Monthly API credits

$100 or $200, "to build on the Claude Platform".

Also on the card

Priority access at high-traffic times, "Higher output limits for all tasks" and "Early access to advanced Claude features".

Read the multipliers slowly. Max 5x "includes five times the Pro plan's per-session usage allowance", Max 20x "20 times", and the session resets every five hours. "Max plans also have a weekly usage limit that applies across all models", and none of the pages I read put a number on it.

Fable is where Max and Pro split. The usage-credits and Claude Code articles add that "Max plans also have a separate weekly limit for Fable." Anthropic stops there. So do I.

How Claude usage limits work

Paid Claude plans run on two clocks, Free on one: "Every plan has usage limits that reset on a rolling five-hour session window, and paid plans add weekly limits on top." The five-hour window is the one you feel during a working day. The weekly one ends a heavy week early, and it "resets at a fixed time each week that is assigned to your account", per the Pro, Max and Team help articles.

Anthropic publishes no message count for any plan

The pricing FAQ says it outright: "there's no fixed message count." Instead it gives multiples, each measured per five-hour session: "Pro gives you at least 5x more usage per 5-hour session than Free. Max gives you 5x or 20x more usage per 5-hour session than Pro. On Team plans, Standard seats give more than Pro and Premium seats give 5x more than Standard." So Max and Team are defined against Pro, and Free's own allowance has no published number at all.

It is one pool, too. "Your activity across Claude on web, desktop, mobile, and Claude Code all draws from the same pool." A long Claude Code session draws on the same allowance your chat uses.

And Anthropic keeps a free hand. It "may limit your usage in other ways, such as weekly and monthly caps or model and feature usage, at our discretion." Any page promising a fixed number of Claude messages is promising more than Anthropic does.

You have three options when you hit a limit

Per the FAQ, you can wait for the reset, move to a higher plan, "or, on paid plans, turn on usage credits to keep working at standard API rates." Free gets no third option.

Usage credits work like a prepaid meter: you add funds first, and you can set a monthly spend limit and switch on auto-reload, per Anthropic's usage-credits article. Set the spend limit before anything else. The API table further down shows the rates you would be paying.

Per dollar, Max 20x buys session usage at half the usual rate

Here is a sum Anthropic leaves to you. Divide each plan's price by its per-session multiple of Pro, and you get what each plan charges a month for one Pro-sized session allowance.

My arithmetic on prices read 13 October 2026, with multiples from the Max and Team help articles. Free is left out because Pro is "at least" 5x Free, a floor with no exact ratio. Team plans need a minimum of two members.

Every monthly plan but one lands on exactly $20, at October 2026 prices. Annual billing trims Pro to $16.67 and both Team seats to $16. Max 20x halves it to $10. As upgrades, Pro to Max 5x costs $80 more and adds four Pro-sized allowances per window; Max 5x to Max 20x costs $100 more and adds fifteen.

Two cautions. This measures the per-session allowance only. Every paid plan also has a weekly limit with no published size, and Max adds a separate weekly limit for Fable, so the table says nothing about how much you can use in a week. And a cheap rate only helps if you use it: someone on Max 20x who works like a Pro subscriber is paying $200 for $20 of need.

Is Claude Code included in Pro and Max?

Yes on every paid plan, no on Free. The pricing FAQ says "Claude Code is included in all paid plans. It shares the same usage limits as the rest of your plan", and the table marks it "No" for Free, as of October 2026. Team seats include it too.

Anthropic's Claude Code article says the plan "also covers Claude Code in supported IDEs, including VS Code, Cursor and other VS Code forks, and JetBrains IDEs". Set an ANTHROPIC_API_KEY environment variable and Claude Code bills at API rates instead of the subscription, the pay-as-you-go route the FAQ points to "For heavy coding sessions".

Max's monthly API credits do not reach "Interactive Claude Code in the terminal, IDE, desktop, or web". For the subscription priced against its rivals, see our Claude Code against Cursor and Copilot pricing table.

Claude Team and Enterprise pricing

Team covers groups of 2 to 150, priced per seat. You can "Mix and match seat types", so Premium goes only to the people who need it.

A Team seat costs $25 or $125 a month, and you can mix the two

As of October 2026, a Standard seat is $25 a month billed monthly or $20 billed annually, and a Premium seat is $125 or $100. The Team help article agrees on all four prices, a "minimum of two members" and "up to 150 seats".

Standard seats include 1.25x Pro's per-session allowance and Premium seats 6.25x, each with a weekly limit. Limits are "per-member, rather than applied to the team as a whole", so a quiet colleague's allowance never tops up a busy one's. For Fable, the card reads "50% of weekly limits* on premium seats". The plan adds SSO, central billing, admin controls for connectors and "No model training on your content by default", and Claude Code is included.

A worked example. Three Standard and two Premium seats cost 3 x $25 + 2 x $125 = $325 a month on monthly billing, or 3 x $20 + 2 x $100 = $260 a month on annual billing ($3,120 a year), at October 2026 prices.

Team API credits are set per seat, pooled and capped at $500

The card says "Up to $500 per month in pooled API credits", and "up to" matters. Anthropic's monthly API credits article sets them at $20 per Standard seat and $100 per Premium seat, as of October 2026, "pooled into one monthly balance, capped at $500." Its own example: "a team with three Standard seats and two Premium seats receives $260 a month."

That is the worked example's annual seat bill, to the dollar. Each seat's credit equals its annual-billing monthly price until the pool reaches $500, at 25 Standard seats or five Premium. Matching numbers, different money. The credits buy Claude API calls, expire every cycle, and don't apply to interactive Claude Code or extra usage.

Enterprise is a $20 seat with a usage bill on top

The Enterprise card reads "Seat price + usage at API rates": "US$20/seat/month, billed annually. Usage cost scales with model and task." So the $20 is where the bill starts, as of October 2026.

You can buy it self-serve or through "Contact sales", which adds invoicing and net terms, trials and tiered incentives on committed spend. Enterprise brings SCIM, audit logs, a Compliance API, custom data retention and IP allowlisting, and the pricing page lists a "HIPAA-ready offering available". The page prints no seat minimum.

Education is priced for a whole institution

The Education plan is "a comprehensive university-wide plan for an institution, including its students, faculty, and staff", with "Comprehensive access for all university members at discounted rates". No price is printed. For students paying their own way, our AI study assistant guide weighs the options, and Pro costs the standard price: both plan help articles say "We do not offer standard discounted pricing for any of our paid plans", though "We occasionally run limited-time promotions".

Claude API pricing, briefly

The API is its own bill. Pro does not include it, and Max and Team include it only as monthly API credits. Here are the four current models on claude.com/pricing, per million tokens (MTok), as of 13 October 2026.

Standard rates with the 5-minute cache, from claude.com/pricing and Anthropic's API docs on 13 October 2026. The two agreed on every figure.

Opus 5.5 is the page's "Daily driver for agentic coding and enterprise work"; Fable 5.1 is "Next generation intelligence for long-running agents". The modifiers, as of October 2026:

• Batch processing: "Save 50% with batch processing".
• US-only inference costs 1.1x for input and output tokens.
• Fast mode for Opus 5.5 promises "up to 2.5x faster speeds" at 2x standard pricing, which the docs list as $8 input and $40 output per MTok.
• Web search: "$10 / 1K searches".
• Code execution: "50 free hours of usage daily per organization" on the pricing page, then "$0.05 per hour per container".
• Managed Agents charge "$0.08 per session-hour for active runtime" on top of standard token rates.

Our Claude API price history tracks how these numbers got here, and you can count a job's tokens with our free token counter before you estimate its cost.

One comparison on the same 13 October page surprised me. Opus 5.5 lists at $4 in and $20 out, while the older Opus 5 and Opus 4.5 to 4.8 still list at $5 and $25. The newest Opus costs less than the Opus models before it.

Which Claude plan should you pick?

Match the plan to the week you actually have. Prices are as of October 2026.

Occasional questions: Free. $0 buys Sonnet and Haiku, web search, file creation, running code and up to five Projects. If the five-hour limit never stops you and you do not need Research, Opus or Claude Code, Free covers it.

Daily writing, research or analysis: Pro. $20 a month adds Research, Opus, Claude Code and at least 5x Free's session usage. Go annual at $200 once you are sure.

Pro's limits stop you most weeks, or Claude Code fills your day: Max 5x. $100 buys five times Pro's per-session allowance, priority access, and $100 a month in API credits if you also build on the platform.

All-day agentic work: Max 20x. At $200 a month it has the lowest monthly price per Pro-sized session allowance in the table above, $10, though the weekly limits behind it are unpublished. Fable gets "50% of weekly limits*" on both Max tiers, where Pro meters it through usage credits.

Between 2 and 150 people who need SSO and one bill: Team. Put everyone on Standard and move only the heaviest people to Premium.

SCIM, audit logs or custom data retention: Enterprise. Budget for both halves: $20 a seat billed annually, plus usage at API rates.

Building Claude into your own product: the API. Pay per token, and send anything that can wait through batch processing at half price.

If ChatGPT is still on your shortlist, our ChatGPT pricing explainer lays out OpenAI's plans, is ChatGPT Plus worth it? takes the single-plan question, and Claude vs ChatGPT workspace agents compares the two on agent work.

FAQ

Sources