Claude Enterprise Pricing: Wetin Dem Really Dey Charge

Wetin Claude Enterprise cost?

Claude Enterprise pricing get two parts, but na only one dem publish. As of October 2026, dem list the seat at $20 per user every month, with annual billing. Dem bill usage separately: every token wey your team spend across Claude surfaces, including Chat and Claude Code, dey charged at standard API (application programming interface) rates on top of the seat. No usage allowance dey bundled inside. The seat line dey fixed by the number of seats wey you buy. Nothing dey fix your usage line.

Na so the bill take be. The next part explain who qualify, wetin the seat add over the Team plan, and how dem calculate the number wey dem quote you.

Why Claude Enterprise pricing get seat price but no list price

Enterprise seat dey buy permission to use Claude under your company controls. E no come with any tokens. No published monthly total dey because two seats fit differ by hundred times: manager wey dey ask four questions per day and engineer wey dey run Claude Code across big repository both cost one seat, but their token bills no resemble at all. This lack of clear pricing no be only Anthropic problem, although e no spread equally: na only one of the two vendors dey publish enterprise seat price at all, and you need know this before you try compare two quotes side by side.

How you go pay depend on the door wey you use. The self-serve path takes card or ACH bank transfer and runs on credits wey you buy upfront, so usage dey reduce balance wey you top up. The sales-assisted path bills usage monthly in arrears. Self-serve payment na card or ACH only, no matter the size. You fit pay sales-assisted invoices by bank transfer, or by card when dem smaller. The $50,000 figure wey you fit don see na payment-method rule, no be eligibility minimum: as of October 2026, invoices of $50,000 or more fit only get payment through bank transfer (ACH or wire). Enterprise dey also sell through AWS Marketplace, with monthly-in-arrears billing like the sales-assisted path. If you already get AWS agreement, ask your account team whether Marketplace spend dey count against your committed spend, because na your AWS contract dey define that answer, no be Anthropic's price list.

How many seats Claude Enterprise dey require?

Minimum dey two. Self-serve Enterprise start for 20 seats. Sales-assisted Enterprise start for 50 seats. Na minimum number of seats wey you buy be this, no be requirement for total staff: smaller team fit buy 20 seats and leave some unused. If you remove one member, the seat wey you buy still dey available, and bill no go reduce. Team cover 2 reach 150 seats, so organisation wey need pass 150 seats go Enterprise because Team don reach limit, no be because of one particular control. The limit count the seats wey you buy, no be the people wey you employ: company wey get 2,000 staff fit put 90 of dem for Claude and still remain for Team.

[
  {
    "plan": "Pro",
    "usd_per_seat_month": 17,
    "min_seats": 1
  },
  {
    "plan": "Team standard",
    "usd_per_seat_month": 20,
    "min_seats": 2
  },
  {
    "plan": "Team premium",
    "usd_per_seat_month": 100,
    "min_seats": 2
  },
  {
    "plan": "Enterprise self-serve",
    "usd_per_seat_month": 20,
    "min_seats": 20
  },
  {
    "plan": "Enterprise sales-assisted",
    "usd_per_seat_month": 20,
    "min_seats": 50
  }
]

When you read that chart, remember one thing. Team premium seat wey cost $100 every month get five times the usage of standard Team seat, and that usage dey included. Enterprise seat wey cost $20 every month get no included usage. Na the smaller number get the open end.

Wetin Enterprise get wey Team no get

Start with wetin Team already get, because na here upgrade argument dey usually go wrong. Team get SSO (single sign-on), domain verification, JIT (just-in-time) provisioning, Owner, Admin and User roles, plus spend controls for both user and organisation level. Single sign-on no be reason to move go Enterprise. E dey there already.

Enterprise add the controls wey security team dey specifically ask for:

• SCIM (system for cross-domain identity management) provisioning, so identity provider fit deprovision person wey leave instead of person doing am by hand.
• Audit logs, wey Anthropic list as Enterprise only. OpenTelemetry monitoring for Claude Cowork no dey that list: as of October 2026, Team get am too, no be Enterprise alone.
• Compliance API wey dey return usage data, activity logs, chat histories and file content.
• Custom data retention controls.
• Domain capture and domain claiming, so existing accounts for your company domain fit enter the organisation. Team fit verify domain and block new personal accounts, but e admins no fit claim or migrate existing ones; members fit migrate their own accounts by choice.
• Network-level access control and IP allowlisting.
• Customer-managed encryption keys. You provision key for your own cloud account, and Anthropic use that key protect your team's chats, projects and files.
• US-only inference, wey keep your company's inference inside United States.
• HIPAA-ready offering (Health Insurance Portability and Accountability Act), plus fine grained role-based access control.

One idea dey run through that list: evidence. Team let person use Claude with sensible default settings. Enterprise let you prove afterwards wetin that person do. When auditor ask which files employee upload last quarter, Compliance API answer am and Team no fit, because Team no expose audit log and no endpoint to query one.

Data retention and model training

By default, dem no dey use content for commercial plans train Anthropic models. This one apply to Team and Enterprise too, so “dem go train with our conversations” no be reason to upgrade. Wetin Enterprise add na control over how long dem go keep content, plus ability to read the content back programmatically.

For im commercial products, Anthropic list SOC 2 Type I and Type II reports, ISO 27001 and ISO/IEC 42001 certification. Im Data Processing Addendum, wey be part of the commercial terms, cover GDPR (General Data Protection Regulation). Na properties of the service, no be line items wey you go negotiate. How long dem go retain data, HIPAA offering, customer-managed keys, and inference region na the parts wey enter contract.

Context window no longer be the reason

Old write-ups dey describe 500K token context window as the main Enterprise feature. Check that claim against the current help center before you buy because of am. As of October 2026, Anthropic documents 1M token context window for Opus 5.5, Opus 5, Sonnet 5.5 and Sonnet 5 when you dey chat with Claude on all paid plans. E documents 500K for Opus 4.8 and Sonnet 4.6. Claude Code supports 1M for those older models too, but Sonnet 4.6 need usage credits except for usage-based Enterprise plans. Context length depend on the model and surface wey you choose, so check both before you buy because you want bigger window. Choosing between dem na separate question, and which Claude model to use for which job cover am.

Wetin dey decide your quote

Four things dey move the amount.

• Number of seats. If una buy less than 20 seats, dem no dey sell Enterprise at all. From 20 to 49, una dey use self-serve path with fixed terms. From 50 and above, una still fit buy self-serve without talking to sales, or go through salesperson. Na the sales-assisted path dey make custom terms possible.
• Term. Usage-based Enterprise dey use one seat type, with price per user per month and annual billing, except organisations wey ready for HIPAA. Dem use separate Chat and Chat + Claude Code seats. Annual commitment na the normal deal structure, no be discount wey you need request. This different from the plans wey una people dey use now, where cancelling takes effect at the end of the billing period you already paid for; so treat the year as decision wey una make once, instead of something una fit reverse for April.
• Expected usage. Usage billing dey follow API rates with no allowance, so na here your risk dey. Estimate am from one real month of Team or API usage, no be from headcount, because headcount no tell you anything about tokens. The model wey your people dey use fit change the amount as much as how much dem type, so price the mix wey you expect: wetin Fable 5 cost per million tokens and the jobs wey e fit handle na the kind per-model number wey suppose enter the estimate.
• The extras. HIPAA readiness, customer-managed encryption keys, US-only inference, and custom retention length na separate requests. Mention dem for the first call, because dem fit change the agreement wey una go sign. US-only inference also change token price: as of October 2026, e dey bill usage at 1.1 times standard API rates for Opus 4.6, Sonnet 4.6, and later models.

Before una make that call, get one real number: the tokens wey your people use last month. Where Claude Code tokens actually dey go and how Claude usage limits dey work both explain how to read that number. Quote wey you no fit check na quote wey you no fit challenge.

Solo, small team, or org

• One person. Pro or Max. Enterprise require make you buy at least 20 seats, even if na only one you dey use. Start from choosing between the Claude plans. If you never sure say paid tier worth the money, decide based on how often the free limit really stops you in the middle of work. If Pro look like the answer, check what Pro costs and where its usage limits stop you before you commit to one year. If ChatGPT subscription dey also your shortlist, the individual tiers from both vendors dey close enough in price, so na the work you do go decide.
• 2 to 150 people on Claude, no audit obligation. Team. You already get single sign-on, domain verification, JIT provisioning and spend caps, with usage included for the seat. The Claude Team plan for a small business cover the seat calculation.
• More than 150 people on Claude, or under an audit obligation. Enterprise. The trigger na either the Team seat cap or one named control from the Enterprise-only list, usually SCIM or audit logs.
• Machine workloads with no human behind them. The API. Nightly batch job no need seat, and giving am one go waste the fee.

When Enterprise make more sense than putting the API on a corporate account

Enterprise usage dey billed with the standard API rates. So if you move team from corporate API account to Enterprise, token cost no go change unless you enable US-only inference. E add fixed fee for every purchased seat. That fee give you identity and audit controls wey raw API key no get.

API-only setup give you metered usage with no seat minimum and no monthly floor. But no floor no mean say e free: API no get permanent free tier. New users get small credit to test am. As of October 2026, researchers wey dey work on qualifying AI safety and alignment topics fit apply to Anthropic's External Researcher Access Program for $1,000 API credits. Apart from those grants, meter dey run for every call. Max and Team subscriptions now include monthly API credits. As of October 2026, Team fit get up to $500 per month pooled. But those credits na subscription benefit, not free tier. Enterprise plans no qualify for dem. API-only setup get its own identity and audit controls: Console support SSO when you enable am through Sales, and SCIM for eligible organisations. Standalone Console organisations fit query Compliance API's Activity Feed. API-key workloads no expose message content through that API. Personal API keys identify one user. Shared workloads fit use service accounts or Workload Identity Federation.

The break-even depend on how many people need supervised access. Multiply purchased seat count by monthly seat fee. Then compare the result with the cost to build the same controls yourself: proxy in front of API, your own key rotation, your own logging, and your own per-user attribution. If na under 20 people, start the comparison with the cost of 20 purchased seats, including seats wey you no use. If na a few hundred people, seat fee normally smaller than the engineering work wey e replace. If those people na engineers wey dey use Claude Code, seat versus token comparison get its own figures. wetin Claude Code seat include for Team and Enterprise explain those figures. First run the same calculation for one person with break-even between Claude API and subscription. The answer get the same pattern, and the figures easier to verify.

FAQ

Claude Enterprise plan dey cost how much for each user?

As of October 2026, published seat price na $20 per user every month, with yearly billing. This price na for the seat only. Every token wey your team use dey billed separately with standard API rates, or 1.1 times those rates for Opus 4.6, Sonnet 4.6 and later models if you turn on US-only inference. No usage allowance dey included. So total cost na seat price times purchased seat count, plus actual usage. No complete fixed price dey published because usage per seat fit vary too much to calculate ahead.

Wetin be the minimum seat count for Claude Enterprise?

Twenty seats if you buy self-serve with card or ACH transfer. Fifty seats if you buy through salesperson. These minimums apply to purchased seats. Team wey get fewer than twenty people fit buy twenty seats and leave some unused. Team itself get maximum of 150 seats. So organisation wey need more than 150 seats go move to Enterprise, whether e need the extra controls or not. This limit count the seats wey you buy, not the total number of employees for your company.

Claude Enterprise include usage, or dem dey bill am separately?

Separately. Enterprise charge every token with standard API rates on top of the seat fee, and no allowance dey bundled inside. Pro, Max and Team include usage with the seat, but usage limits dey apply. You fit also buy paid usage credits beyond those limits. Self-serve Enterprise dey use credits wey you buy upfront, while sales-assisted Enterprise dey send invoice monthly after usage.

I need Enterprise to get SSO for Claude?

No. Team already include single sign-on, domain verification and just-in-time provisioning. Wetin Team no get na SCIM provisioning and audit logs. Those two na usually the reason organisations move to Enterprise, together with Compliance API wey dey read the data back. If the requirement na for employees to sign in with their work identity, Team meet am.

Company data dey used to train Claude for Enterprise plan?

No. By default, content for commercial plans no dey used to train Anthropic's models. This rule apply to Team just as e apply to Enterprise. Enterprise add custom data retention controls, Compliance API for exporting activity logs, chat histories and file content, customer-managed encryption keys, and US-only inference option.