Claude has seven different price points, and the one most people quote is not the cheapest.
The Claude price most people repeat is $20 a month, but the cheapest paid seat is $17 and the cheapest team costs $40.
Here is what every plan costs as of September 2026, taken from Anthropic's own pricing pages.
Key Takeaways
Claude runs from $0 to $200 a month for one person, with Pro at $20 billed monthly or $17 billed annually, and Max at $100 or $200.
Team seats cost $20 or $100 per seat per month on annual billing with a two seat minimum, while Enterprise is $20 per seat plus usage charged at API rates.
Anthropic publishes usage multipliers instead of message counts, so a fixed prompt count is an estimate rather than a published figure.
Claude Code, Claude Cowork and Claude Design carry no separate price and come with every paid plan.
Sticker prices have not increased, but Fable model access moved from included to metered on Pro and Team standard seats in July.
No paid plan has a student discount, and Claude for Education is quoted per institution rather than published as a price.
Claude costs $0 on the Free plan, $20 a month on Pro or $17 a month with annual billing, $100 a month for Max 5x and $200 for Max 20x.
Business plans are charged per seat: $25 or $20 for a Team standard seat, $125 or $100 for a Team premium seat, and $20 per seat plus metered usage on Enterprise.
That is five named plans across seven price points.
Prices are for US customers and exclude tax.
Plan | Billed monthly | Billed annually | Seats | Usage per 5-hour session |
Free | $0 | $0 | 1 | Baseline |
Pro | $20 | $17 ($200 up front) | 1 | At least 5x Free |
Max 5x | $100 | Monthly only | 1 | 5x Pro |
Max 20x | $200 | Monthly only | 1 | 20x Pro |
Team standard seat | $25 per seat | $20 per seat | 2 to 150 | 1.25x Pro |
Team premium seat | $125 per seat | $100 per seat | 2 to 150 | 6.25x Pro |
Enterprise | $20 per seat plus usage at API rates | Annual billing only | Custom | Consumption based |
Two numbers in that table matter more than the headline prices.
Annual billing on Pro saves $40 a year, and Anthropic charges the full $200 up front rather than spreading it.
The real floor for a company is not $20 either, because Team needs at least two members, which puts the cheapest possible team at $40 a month on annual billing.
The Free plan covers chat on web, iOS, Android and desktop, web search, memory across conversations, file creation with code execution, desktop extensions, Slack and Google Workspace connectors and extended thinking.
Pro adds more usage plus four products the Free plan cannot open at all: Claude Code, Claude Cowork, Claude Design and Claude Science.
Pro also unlocks unlimited projects, Research, Claude in Chrome, Claude for Microsoft 365 and access to more models.
The gap is wider than the price suggests.
Free users get Sonnet and Haiku.
Pro users get Opus as well, the model Anthropic positions for complex agentic coding and enterprise work.
One thing Pro does not add is collaboration.
Project sharing and the @Claude assistant stay locked until you move to a Team plan, so several Pro seats do not add up to a shared workspace.
Max costs $100 a month for the 5x tier and $200 a month for the 20x tier, and Anthropic sells both as usage upgrades rather than feature upgrades.
Every Max subscriber gets the same products a Pro subscriber gets, plus higher output limits, early access to new features and priority access when traffic is heavy.
Anthropic's Max plan support article states the prices are for web subscriptions and that mobile app store pricing may differ. Worth knowing before you compare Max against Pro's annual rate: that same article says Max is currently sold as a monthly subscription only, while the comparison table on the pricing page still lists monthly and annual billing for both Max tiers.
The pricing page FAQ sides with the support article and says both Max options are billed monthly.
Treat Max as a monthly plan until Anthropic makes the two pages agree.
One detail that saves money on the switch: if you move from an annual Pro plan to Max and your remaining Pro balance is larger than the Max price, Anthropic credits the balance to your account, as long as the billing address matches.
Context length is not a reason to buy Max.
The context window stays at 200k on Free, Pro, Max and Team, and only Enterprise raises it to 500k on the default model.
Claude Code has no price of its own.
It is included in every paid Claude plan, and it draws from the same usage pool as your chats, so terminal work and conversations spend the same allowance.
This is where older guidance no longer matches the current documentation.
The rule that Claude Code needs a premium seat now applies only to older Enterprise contracts, where access sits on Chat plus Claude Code seats or premium seats.
New and self-serve Enterprise plans include it with the standard seat.
Your seat also covers Claude Code inside VS Code, Cursor and other VS Code forks, and JetBrains editors such as IntelliJ and PyCharm, billed exactly like terminal use.
Usage-based Enterprise plans skip per-seat limits entirely and bill consumption at API rates.
Model access, not price, is the real difference between Claude's tiers as of September 2026.
Sonnet and Haiku are on every plan including Free, Opus starts at Pro, Fable is handled differently on almost every tier, and Mythos is not on any consumer plan.
Model | Free | Pro | Max 5x and 20x | Team | Enterprise |
Haiku | Yes | Yes | Yes | Yes | Yes |
Sonnet | Yes | Yes | Yes | Yes | Yes |
Opus | No | Yes | Yes | Yes | Yes |
Fable | No | Usage credits | Included, up to 50% of weekly limits | See below | Yes |
Mythos | No | No | No | No | Limited availability |
Fable is the tier worth reading twice.
Anthropic's Fable models support article explains that on Max plans and premium seats, Fable 5 and Fable 5.1 are a standard part of the plan and you can spend up to half your weekly limit on them at no extra cost. On Pro plans and Team standard seats, both Fable models run on pay-as-you-go usage credits instead.
Fable also draws down weekly limits faster than other Claude models, which matters most on the plans where it runs on credits.
Here the official pages disagree again.
The comparison table on the pricing page marks Fable as unavailable on Team, while the Fable support article says Fable 5 and 5.1 are available on all paid plans and describes exactly how Team standard and premium seats each get them.
The support article is the more specific and more recent of the two, so that is the one to plan around.
Anthropic's Team plan support article sets the minimum at two members and the ceiling at 150 seats, above which you move to Enterprise. The same page publishes the usage multipliers that make the seat types make sense.
A standard seat gives 1.25x more usage per session than Pro.
A premium seat gives 6.25x.
You can mix both on one bill, which is the point: premium seats for the heaviest Claude Code users and standard seats for everyone else.
Usage limits are per member rather than pooled, so one heavy user cannot drain the team's allowance.
Team plans buy administration as much as capacity.
Single sign-on, domain verification, spend controls, enterprise search across Slack and Microsoft 365, central billing and no model training on your content by default all arrive at the Team tier, while the pricing page reserves domain capture and role-based access for Enterprise.
Enterprise is priced differently on purpose.
The pricing page lists $20 per seat per month plus usage billed at API rates, with annual billing, and adds SCIM, audit logs, a Compliance API, custom data retention, network access control, IP allowlisting and a HIPAA-ready offering.
You can start Enterprise self-serve or ask for a tailored quote, and Enterprise is the only tier that takes ACH, with net-term invoicing reserved for sales-assisted agreements.
Neither product has its own price.
What they cost you is usage.
Cowork and Claude Code both draw on the same plan allowance as your chats, and when that allowance runs out, usage credits let team members keep working in Claude, Cowork and Claude Code at standard API rates.
That makes the honest answer to "how much does Claude Cowork cost" a question about your own workload rather than a number.
Hand Cowork long multi-step jobs on a $17 Pro seat and you can reach your weekly limit, at which point credits start.
The same work on a $100 Max seat may never touch them.
There is no student price for Claude and no discount code for any paid plan.
Anthropic states plainly in its Max plan article that it does not offer standard discounted pricing on any of its paid plans, that promotions are occasional and limited, and that its support team cannot issue one-off discounts or coupons.
Students have two legitimate routes instead.
The first is the Free plan, which anyone in a supported location can use with an email address.
The second is institutional: Anthropic's higher education page describes a university-wide plan covering students, faculty and staff at discounted rates, with dedicated API credits and a learning mode, arranged through its education team. No dollar figure is published for it, because it is quoted per institution.
Researchers get a better deal than students do.
No Claude plan carries a higher sticker price today than when it launched: Pro has been $20 a month, Max has been $100 and $200 since April 2025, and a Team seat now costs less than the $30 per user Anthropic announced in 2024.
What changed is metering, and that is why so many users feel like they are paying more.
Start with the price that did not go up.
Anthropic's platform documentation records that the $2 and $10 per million token rate for Sonnet 5, launched as introductory pricing through the end of August, is now the standard price, and that the scheduled increase to $3 and $15 on September 1 will not happen.
Compare that with Sonnet 4.6 at $3 and $15, and the newer model is a third cheaper.
Opus tells the same story: Opus 4.1 cost $15 and $75 per million tokens, while Opus 5 costs $5 and $25.
Now the change that did cost people money.
Anthropic ran a promotion that let subscribers spend up to half their weekly limit on Fable 5 at no extra cost, and that promotion ended on July 19, 2026.
It covered Fable 5 only, and Fable 5.1 was never part of it.
Pro subscribers and Team standard seats now pay for Fable usage with credits, while Max plans and premium seats keep it inside their weekly limits.
Eligible Pro and Team standard seats received a one-time credit when the change landed, with no equivalent credit for Fable 5.1.
There is a third, quieter shift.
Anthropic's documentation notes that Claude 4.7 and later models use a newer tokenizer that produces roughly 30% more tokens for the same text.
On a subscription that shows up as limits arriving sooner.
On the API it shows up as a bigger bill at an unchanged rate.
The Claude API has no monthly fee and bills per million tokens, which is also where subscription overflow gets charged once usage credits are on.
Rates below are the published prices as of September 2026.
Model | Input per million tokens | Output per million tokens |
Fable 5.1 | $10 | $50 |
Opus 5 | $5 | $25 |
Sonnet 5 | $2 | $10 |
Haiku 4.5 | $1 | $5 |
Three discounts change the real cost.
Batch processing takes 50% off both input and output for work that can wait.
Prompt caching charges cache reads at a tenth of the input price, dropping to 2.5% on Fable 5.1, with cache writes at 1.25x for five minutes or 2x for an hour.
The full 1M token context window is included at standard pricing on Claude 4.6 and later, so a 900,000 token request costs the same per token as a small one.
Anthropic's own worked example is the clearest way to see the scale.
A one-hour coding session on Opus 5 using 50,000 input and 15,000 output tokens comes to $0.705, made up of $0.25 of input, $0.375 of output and $0.08 of session runtime.
With prompt caching active on 40,000 of those input tokens, the same session drops to $0.525.
Separate rates apply to tools rather than models: web search is $10 per 1,000 searches, code execution is free for the first 1,550 hours a month per organization and then $0.05 per container hour, and US-only inference carries a 1.1x multiplier.
No, and Anthropic says so without publishing the local numbers.
Every figure on the pricing page is in US dollars and excludes tax.
The Team plan article is more explicit: the prices shown are for US customers, and pricing, currency and tax handling vary by region.
Anthropic points you to the upgrade page inside your own account to see the price for your location, so the official figure in rupees or pesos appears only at your own checkout.
Two practical consequences follow.
Any local-currency price you find on a third-party site is unofficial unless it came from your own checkout screen.
And the platform you subscribe through matters, because the Max plan article notes that its prices cover web subscriptions while mobile app store pricing may differ.
Read the 2026 changes together and a strategy shows through.
Anthropic left its headline seat prices where they were, priced its newest volume model below the model it replaced, and let the promotion that included its most expensive model expire into metered credits.
That combination reads as revenue growth coming from usage rather than from price rises.
Seat prices are the number customers compare and complain about, so they stay still.
Metering is where a heavy user's cost shows up, and the weekly limit, the usage credit and the 5-hour session window have all become more precise while $20 has stayed $20. Anthropic describes the newer tokenizer as part of what makes those models better, and the arithmetic effect is that the same text now counts as more tokens.
For anyone watching Anthropic as a business rather than as a tool, that is the useful signal from a pricing page.
Research firm Sacra estimates that enterprise and startup API calls drive roughly 80% of Anthropic's revenue, which leaves consumer subscriptions well down the list, so the plans most readers buy are not where most of the company's money comes from.
How much does Claude cost per month?
Claude costs $0 on Free, $20 on Pro or $17 with annual billing, $100 on Max 5x, $200 on Max 20x, $25 or $125 per Team seat, and $20 per Enterprise seat plus usage.
Is Claude free to use?
Yes, the Free plan costs nothing and includes chat, web search, memory, file creation and connectors, but not Claude Code, Cowork, Design or Opus.
Is Claude Code included in my plan?
Yes, Claude Code is included in every paid plan and shares one usage pool with your chats, with usage credits or a Console account covering anything beyond the plan.
What is the difference between Claude Pro and Max?
Max costs five to ten times more than Pro and buys 5x or 20x the session usage, higher output limits and priority access, not extra products.
How much is a Claude Team seat?
A standard seat is $25 a month or $20 on annual billing and a premium seat is $125 or $100, with a two seat minimum and up to 150 seats.
Does Claude have a student discount?
No, Anthropic does not discount any paid plan for students, and the only education pricing is a university-wide plan quoted through its education team.
Did Claude increase its prices?
No, the sticker prices are unchanged and Sonnet's per-token price fell, but Fable access moved to metered usage credits on Pro and Team standard seats in July.
How many messages do I get on each plan?
Anthropic does not publish a message count, because limits depend on conversation length, model choice and features, and it describes capacity as multipliers with a 5-hour session window plus a weekly cap.
Can I buy shares in Claude or Anthropic?
Claude is a product rather than a company, and Anthropic is still privately held, so its shares are not listed on a stock exchange; pre-IPO perpetual futures that track its implied valuation are derivatives rather than equity.
Pick the plan by how much you use Claude, not by which features you want, because almost every paid tier now carries the same products.
Pro at $17 on annual billing is the value pick, Max is a usage purchase, and Team only makes sense once you need shared projects and admin control.
Check the figures against Anthropic's own pages before you buy, since it reserves the right to change them.
Then, if the business behind the pricing interests you more than the subscription, that is a different story worth reading.