
What Claude costs in 2026: Free, Pro at $20, Max at $100 or $200, Team and Enterprise, the five hour and weekly limits, Claude Code sharing, and API rates per million tokens. Checked November 6, 2026.
Claude's current plans include Free, Pro, Max, Ultra, Team, and Enterprise. Pro is $20 monthly or $17 per month with a $200 annual payment. Max is $100 per month for 5x Pro usage; Ultra is $200 per month for 10x weekly Pro usage on verified accounts. The official page does not list a current Max 20x tier, Team covers 2 to 150 seats and Enterprise is priced on request. API usage is billed separately per million tokens: Opus 5.5 is $4 input and $20 output, Sonnet 5.5 is $2 and $10, and Haiku 5.5 is $0.10 and $0.50 for prompts up to 100K tokens. Subscription limits vary with use and include session and weekly limits, not a fixed number of messages.

For the next decision after claude pricing 2026, use ChatGPT Business vs Claude Team: Seats, Admin and Data Ownership, and Team AI Plan vs Individual Subscriptions: Who Owns the Work?, and ChatGPT Enterprise Alternatives for Mid-Size Companies: A Requirements Shortlist. These business guides focus on the evidence, ownership and buying questions behind the choice.
For the corresponding account, billing or archive decision, read ChatGPT + Claude + Gemini Cost: Monthly and Yearly, Move Gemini History to Claude: Takeout and Krater Import.
Anthropic sells individual subscriptions, organization plans, and API access. These are separate billing routes. The U.S. prices below were checked on November 6, 2026; taxes and regional availability may change what appears at checkout. See Anthropic's current pricing page and the Team plan help page.
Households weighing one login against separate subscriptions can read Claude family plan rules and costs before choosing a plan.
Regional checkout prices can change the personal-plan decision; whether Claude Pro is worth paying for weighs Claude Pro against actual use.
To connect an individual Claude subscription with organization plans and API billing, see Claude Team and Enterprise for business.
The price of an individual subscription does not include a fixed API token balance. Team seat prices also do not mean unlimited requests; member-level usage policies apply. For Enterprise, use Anthropic's sales route rather than inferring a seat price from Team. The plan table is a snapshot, not a promise that future prices will stay the same.
Claude Code can be used through eligible Claude paid plans or through API-based workflows, and those paths have different bills. Anthropic's help documentation says Claude and Claude Code activity share usage limits on Pro and Max; Team documentation says Claude Code is included with every Team seat, and new or self-serve Enterprise seats also include it. A plan that includes Claude Code still has usage limits.
| How you use Claude Code | Billing route | What to watch |
|---|---|---|
| Interactive coding under an individual Claude account | Pro, Max, or Ultra subscription | Shared Claude app and Code usage limits. |
| Managed organization workspace | Team or eligible Enterprise seat | Seat type, per-member limits, and administrator controls. |
| Automation using an API key | Anthropic Platform API | Model token rates, caching, Batch, tools, and budget alerts. |
For a solo developer, Pro at $20 a month is a reasonable first measurement point if interactive coding is the main use. If weekly work repeatedly hits the visible limit, compare Max at $100 for 5x usage or Ultra at $200 for 10x weekly usage. Do not multiply a guessed number of prompts to predict capacity. If the developer runs a service or unattended job, use API billing and calculate from actual tokens instead of treating a chat subscription as an application license.
The current public plan comparison names the $100 monthly tier Max and describes it as 5x Pro usage. The $200 monthly tier is called Ultra and describes 10x Pro usage every week. Those multipliers compare plan usage; they do not promise a fixed number of messages. Anthropic's pricing page does not list a Max 20x tier. See What happened to Max 20x? below.
Both paid tiers are still subject to usage policies. Long conversations, complex model choices, files, and tool calls can consume allowance at different rates. Use Anthropic's Usage page to see the limits that apply to your account and read the separate Claude limits guide for reset details.
Claude Pro is Anthropic's paid individual subscription. The U.S. pricing page lists $20 for monthly billing or $200 for an annual subscription paid up front. The annual amount is not twelve monthly installments; over twelve months, it works out to about $16.67 per month before tax. Pro adds more Claude app usage than Free and includes features shown on the live plan page, including Claude Code. The exact feature bundle can change, so confirm the page before purchase.
Pro is not an unlimited plan. Anthropic describes a five-hour session limit and a weekly limit for Pro activity. Model, conversation length, uploaded material, tools, and activity in Claude Code can affect how quickly the allowance is used. The account Usage page is more useful than a generic “messages per day” estimate because Anthropic does not publish one fixed message count for all users.
Choose monthly billing when you want to test Pro for a short course, project, or work period. Twelve monthly payments would total $240, so the $200 annual option is $40 less if you use it for a full year. If you expect to need it for only a few months, monthly billing can cost less overall despite its higher month-to-month rate. Before subscribing, check renewal date, tax, account region, and how to cancel.
Max is worth evaluating when Pro's limits repeatedly interrupt work that has a real cost: a developer's coding session, a creator's long research sprint, or a consultant preparing several source-heavy deliverables. It is not automatically worth it because it is the larger plan. If usage warnings occur only occasionally, first shorten threads, reduce repeated context, and check whether the work can wait for the reset. If the account rarely reaches a limit, additional capacity may go unused.
Use a two-week log: note when a limit appears, what you were doing, which model was active, and whether waiting for the reset disrupted a deadline. Then compare that evidence with the official difference: Max is listed at 5x Pro usage for $100 monthly, while Ultra is listed at 10x weekly usage for $200. The price is recurring, and the multiplier is not a guaranteed message count. Avoid upgrading based on a hypothetical heavy week that is not representative.
The phrase “Claude Max 20x” still appears in searches, but the current Anthropic pricing page checked for this guide names the $100 tier Max and lists 5x Pro usage. It names the $200 tier Ultra and lists 10x Pro usage every week. The page does not list a current Max 20x plan. So this guide treats Ultra, at $200 for 10x weekly Pro usage, as the current top individual tier. If Anthropic changes the plan names or publishes a new tier, use the official page and its effective date.
Claude Team is a seat-based organization subscription with Standard and Premium seats. Anthropic's current U.S. help page lists Standard at $25 per member per month with monthly billing or $20 with annual billing. Premium is $125 monthly or $100 annually per member. Team requires at least two members and supports up to 150 seats. The help page describes Standard at 1.25x Pro per-session usage and Premium at 6.25x. Limits apply per member, not as one allowance for the whole organization.
Claude Code is included with every Team seat according to Anthropic's help page. Enterprise is intended for larger or more controlled deployments; the public price used here is not a standard self-serve consumer price. Request a quote and check minimums, payment terms, retention controls, identity management, and usage billing directly with Anthropic. Do not treat a Team Standard or Premium rate as an Enterprise quote.
For plan-specific features and comparisons, see Claude Team plan pricing and Claude Enterprise pricing. Both pages should be read alongside Anthropic's current plan terms, because seat availability and usage rules can change.
Anthropic's help page lists Standard at $25 per seat monthly or $20 per seat per month billed annually. Premium is $125 monthly or $100 annually. For ten seats, the arithmetic is:
| Ten-seat mix | Monthly billing | Annual-billing monthly equivalent |
|---|---|---|
| 10 Standard | 10 × $25 = $250 per month | 10 × $20 = $200 per month, or $2,400 for the year |
| 10 Premium | 10 × $125 = $1,250 per month | 10 × $100 = $1,000 per month, or $12,000 for the year |
| 8 Standard + 2 Premium | 8 × $25 + 2 × $125 = $450 per month | 8 × $20 + 2 × $100 = $360 per month, or $4,320 for the year |
These are seat-fee calculations before taxes and optional usage purchases. Team requires at least two members and supports up to 150 seats. Usage limits are per member, so the team should assign Premium based on work patterns rather than assume one shared team allowance. See the detailed Team pricing guide.
Anthropic bills the API by model and token category. The table below shows the current first-party U.S. rates read from Anthropic's API pricing documentation on November 6, 2026. MTok means one million tokens. Batch rates are 50% of standard input and output rates where Batch is available; prompt caching rates are separate and can combine with Batch under the published rules.
| Model | Input / output per MTok | 5-minute cache write | 1-hour cache write | Cache hit | Batch input / output |
|---|---|---|---|---|---|
| Claude Opus 5.5 | $4 / $20 | $5 | $8 | $0.20 | $2 / $10 |
| Claude Sonnet 5.5 | $2 / $10 | $2.50 | $4 | $0.10 | $1 / $5 |
| Claude Haiku 5.5, prompts up to 100K tokens | $0.10 / $0.50 | $0.125 | $0.20 | $0.01 | $0.05 / $0.25 |
| Claude Haiku 5.5, prompts over 100K tokens | $0.50 / $2.50 | $0.625 | $1 | $0.05 | $0.25 / $1.25 |
The Haiku threshold is based on the prompt length of an individual request, including cache reads and writes. A total day of one million tokens does not automatically trigger the over-100K rate if those tokens are split across smaller prompts. Cache writes are billed when content is stored; cache hits are billed when a later request reads it. The table is not a complete application invoice: server-side tools, grounding, media, and other features can have additional charges. Confirm the live table before deployment.
Assume an application sends 1,000,000 input tokens and receives 250,000 output tokens each day, uses no caching or tools, and runs for 30 days. The model-specific rates below are from Anthropic's current API pricing table.
| Model | Daily calculation | Daily cost | 30-day estimate |
|---|---|---|---|
| Claude Opus 5.5 | 1 × $4 input + 0.25 × $20 output | $9.00 | $270.00 |
| Claude Sonnet 5.5 | 1 × $2 input + 0.25 × $10 output | $4.50 | $135.00 |
| Claude Haiku 5.5, prompts under 100K each | 1 × $0.10 input + 0.25 × $0.50 output | $0.225 | $6.75 |
This is a token-only estimate, not a quote. It assumes Haiku requests stay at or below 100,000 prompt tokens each, and excludes server-side tools, extra context, cache storage, and taxes. Batch processing halves standard input and output rates where supported. If your application repeats a large prefix, estimate 5-minute or 1-hour cache writes and subsequent cache hits separately; then monitor actual invoices.
Thirty minutes with Adam Johansson, an AI expert who maps out what actually fits your business: the right tools, plan and setup for your team size and usage. Free, no strings attached.
Free for any business, any size. We reply within one business day.

Adam Johansson
AI Advisor, Krater
Subscriptions and API usage answer different questions. A person who chats in Claude or uses Claude Code in a plan can budget a recurring subscription, then monitor the plan's usage limits. A developer who sends requests from an application pays token rates, with separate math for input, output, caching, Batch, and tools. A Pro subscription does not include a million API tokens per day.
For a solo developer, begin with the work pattern. If Claude Code is used interactively for repository work and the included plan allowance is enough, Pro's listed $20 monthly price is the simplest test. If the developer repeatedly reaches the account's session or weekly limits, compare Max at $100 for 5x usage or Ultra at $200 for 10x weekly usage. Those are relative capacity descriptions, not a promise that every coding task fits. If the developer is building an application, calculate the API bill from the expected tokens instead of upgrading a subscription to cover server traffic.
For an API estimate, count input and output separately, note prompt size and repeated context, then apply current cache or Batch rates only when the request uses those features. A workload with long prompts can have materially different costs from one with short prompts even if both produce the same number of answers. Keep a billing alert and review actual usage after the first week.
Anthropic's Cowork help page describes Cowork as available on paid plans including Pro, Max, Ultra, Team, and Enterprise. The page does not present Cowork as a separate subscription tier with its own standalone price. Treat Cowork access as a paid-plan feature, and confirm the account's current plan eligibility and product limits in Anthropic's help center. Do not add a separate Cowork fee unless the official checkout shows one.
Anthropic's Claude for Education page describes a university-wide plan for an institution, including students, faculty, and staff. It does not publish a universal individual student checkout price. Ask the university whether it has provisioned access and what data rules apply before paying for a personal subscription. For a student-focused comparison of Claude Free, campus access, and Pro, see Claude Pro student discount.
Students who use Claude only during a term may prefer to compare monthly Pro billing with campus access rather than assume an annual plan is always the best value. The $200 annual price equals ten monthly payments at $20; a shorter period of occasional use can cost less on monthly billing. Verify current terms before subscribing.
Krater Pro is a separate multi-model workspace. Its code-backed configuration lists $20 per month or $200 per year and 1,500 monthly credits. The Pro plan page describes 500+ models, including Claude, GPT, and Gemini, plus Krater Agent. Credit use varies with the selected model and task, so this article does not invent a per-answer amount or treat Krater credits as Anthropic API tokens.
Use the in-product credit estimate and balance for the model you select. If your application needs a direct Anthropic API, use Anthropic's Platform billing and the API table above. If you want several model families and agent workflows in one workspace, compare Krater's plan features with your team workflow. These are separate products and billing meters.
Choose by workload and account boundary rather than by a single headline price. The table is a starting point; compare the official plan page and your account's actual limits before changing subscriptions.
| Persona | Plan to evaluate first | Decision check |
|---|---|---|
| Student | Free or a university-provided Claude for Education account; Pro if personal paid use recurs | Ask the institution about managed access and read Claude student discount details. |
| Solo creator | Free, then Pro for regular app work | Compare monthly use, file workflows, and renewal date. |
| Developer | Pro for interactive Claude Code; Max or Ultra if plan limits repeatedly interrupt work | Use API pricing instead for an application or automation. |
| 10-person team | Team Standard for most seats; Premium for members needing higher allowance | Compare per-seat totals and each member's actual usage. |
| Enterprise | Request Anthropic's current Enterprise terms | Include identity, retention, support, usage billing, and contract requirements. |
For a team that uses different model families, a multi-model workspace may also be worth evaluating, but do not assume it replaces Claude's organization controls or API access.
Anthropic lists Claude Pro at $20 per month or $200 for an annual subscription in the U.S. Check checkout for region and tax.
The current public page lists Max starting at $100 per month for 5x Pro usage. The $200 tier is Ultra, described as 10x weekly Pro usage.
Anthropic's current pricing page does not list a Max 20x tier. It lists Max at 5x and Ultra at 10x. Do not treat an older or third-party label as a current official plan.
Pro is the individual baseline paid plan. Max starts at $100 per month and is described as 5x Pro usage. Neither description promises unlimited requests.
Anthropic documents Claude Code access with eligible individual paid plans, Team seats, and new or self-serve Enterprise seats. The exact account and usage terms are on the linked support pages.
Interactive use through a Claude plan draws on that plan's limits. API-key use is separate token-based Platform billing. Check which account the CLI is configured to use.
Start with the plan subscription if using Claude Code interactively. The current Pro, Max, and Ultra prices are above; use API rates for automated application traffic.
Anthropic's help page says Cowork is available on paid plans including Pro, Max, Ultra, Team, and Enterprise. It does not list a separate Cowork subscription price.
Claude Opus 5.5 is $4 per million input tokens and $20 per million output tokens before cache or Batch rates. Sonnet 5.5 and Haiku 5.5 have different rates shown in the API table.
There is no single price because input and output are priced separately and model rates differ. For example, Opus 5.5 costs $4 per million input and $20 per million output tokens.
Do not assume so. Claude app subscriptions and Anthropic Platform API usage are separate billing routes; check the account-specific terms for any current credit offer.
Standard seats are $25 per member monthly or $20 annually. Premium seats are $125 monthly or $100 annually. Team has a two-seat minimum and a 150-seat cap per the help page.
The public plan page does not provide a standard self-serve Enterprise checkout price in the facts used here. Ask Anthropic for current terms rather than estimating from Team.
Anthropic says Batch usage is billed at 50% of standard API prices. Prompt caching can also reduce repeated-context cost; verify current model-specific rates and stacking rules.
Claude pricing has separate lanes for app subscriptions, team seats, Enterprise, and API traffic. Free and Pro serve individual app use; Max and Ultra increase usage capacity; Team separates Standard and Premium seats; Enterprise requires a direct quote; API use is charged by model and token. The best choice depends on workload, limits, administration, and where the data is allowed to go. Check Anthropic's live terms before purchase, and use the separate usage limits guide when the main question is resets rather than price.