Claude Enterprise: What It Costs, What Only Enterprise Has, and the Team Plan Alternative
Claude Enterprise is Anthropic’s plan for large organisations: $20 a seat a month, billed annually, from 20 seats, with every message billed on top at API rates. It adds SCIM, audit logs, a compliance API, custom data retention, role-based permissions and a HIPAA-ready offering to everything in Team.
Many companies that look at Enterprise want control more than those features. They can get it by staying on Team plans, with usage included in the seat, and putting the control in a layer outside Claude.
What Claude Enterprise is
Claude Enterprise is the top business plan for Claude. People get the same Claude as on the Team plan: chat, projects, connectors, skills, Claude Code, Cowork and the rest. What changes is how it is billed and how much the company can control and audit. Team is for 2 to 150 people. Enterprise starts at 20 seats and has no upper limit listed.
Both plans run on Anthropic’s business terms, under which your content is not used to train models by default. If that is the main worry, Team already covers it; see is your data safe in Claude.
Claude Enterprise pricing and billing
From Anthropic’s pricing page, checked 3 November 2026, in US dollars before tax:
| Team | Enterprise (self-serve) | Enterprise (sales-assisted) | |
|---|---|---|---|
| Size | 2 to 150 people | From 20 seats | Agreed with sales |
| Seat price | Standard $20 a month annual, $25 monthly. Premium $100 annual, $125 monthly | $20 a seat a month, annual only | Not listed |
| Usage | Included in the seat, up to its limit | Billed on top at API rates | Agreed with sales |
| Payment | Card only | Card or ACH | Card, ACH or invoice with net terms |
The seat price is the same as Team’s Standard seat, but it buys something different. On Enterprise the seat is access, and the work is metered. Anthropic’s Enterprise consumption guide explains how: usage is counted in tokens and drawn from one consumption pool shared by the whole organisation. The model, the effort level and the length of the conversation all drive cost, and Claude Code and Cowork use considerably more tokens than ordinary chat. Admins can set spend limits for the organisation, for groups and for individual users, and the limits reset on the first of each month.
Team is a fixed monthly bill. Enterprise is a seat fee plus a meter, and the meter runs faster as people get better at using Claude for long documents, research and agents. Our Claude Team vs Enterprise guide has a worked example for a 60-person company and a calculator for your own numbers.
What only Enterprise has
Team already includes single sign-on, domain verification, admin controls for connectors, usage analytics, spend controls, org-wide skills and organisation instructions. According to the same pricing page, Enterprise adds:
- Role-based access with fine-grained permissions. Custom roles decide which features, models and connectors each role may use, down to single tools on a connector (Anthropic’s guide).
- SCIM, so users are created and removed from your directory.
- Audit logs and a compliance API for your own security and compliance tools.
- Custom data retention controls for chats and files.
- Network-level access control and IP allowlisting.
- A HIPAA-ready offering for health data.
- Domain capture and account migration, so accounts on your domain are brought into the company plan, plus Claude Security (beta) and inference hooks.
What Enterprise does not give you
Every control in Enterprise is a control inside Claude. That is what it is built for, and it does it well. It stops at Claude’s edge:
| What companies often want | Claude Enterprise |
|---|---|
| The same rules for which systems the AI reaches, in Claude, ChatGPT and Copilot | Roles control connectors in Claude only. Other AI tools are set up separately |
| One log of what the AI did in company systems, across every AI tool | Claude’s audit logs cover Claude |
| Company skills and rules that follow people between tools | Org-wide skills and instructions apply in Claude |
| The same control for a company that splits Claude across several plans | Each plan is configured on its own |
Few companies of 50 to 1,000 people use only one AI tool. Sales may live in ChatGPT, the developers in Claude Code, finance in Copilot. If the rules for your CRM and finance system have to be written three times, they drift. That is the problem an MCP gateway solves, and why we treat it as a separate layer from the AI plan.
The Claude Enterprise alternative: Team plans plus a control layer
Many companies get what they wanted from Enterprise for less by structuring their Claude use as Team plans and putting the control outside Claude. Anthropic allows this: its account management FAQ answers “Can I create multiple teams?” with “Yes, you can create multiple teams by purchasing multiple Team plans.” Some companies therefore run one Team plan for the tech team and one for the rest of the company, each with usage included in the seat.
Take a company of 220 people: 70 in product and engineering, 150 everywhere else. That is too many for one Team plan. On annual prices from the pricing page, checked 3 November 2026:
| Two Team plans | Enterprise (self-serve) | |
|---|---|---|
| Tech team, 70 people | 50 Standard and 20 Premium seats: $3,000 a month, usage included | 70 seats: $1,400 a month, plus their usage |
| Rest of the company, 150 people | 150 Standard seats: $3,000 a month, usage included | 150 seats: $3,000 a month, plus their usage |
| Every month | $6,000, the same every month | $4,400 plus usage at API rates |
Enterprise only comes out cheaper if the 220 people together use less than $1,600 of Claude a month at API rates, about $7 a person. For a company that uses Claude every day, and a tech team running Claude Code, the fixed Team bill is usually the safer budget.
Two plans are two separate teams, each with its own members, billing and admin settings. That is where the control layer comes in. Deployed OS is connected once and the same MCP connection is added to every Team plan, and to ChatGPT and Copilot if you use them:
- Which systems the AI reaches, per group. User groups decide who may read or change what in each system, and each person signs in with their own login where the system allows it.
- Company context and skills. Shared ways of working that are the same in both Team plans and in the other AI tools.
- Security rules. Guardrails that come with every question, whichever plan or tool it is asked in.
- One log. Every call into your systems is logged, and the log can be exported to your own tools.
The result is Enterprise-style control over your company systems at Team-plan cost. It does not add Claude’s own SCIM, retention settings or compliance API, so if you need those, read the next section. If you are under 150 people, the ordinary route is simpler still: one Team plan plus Deployed OS. Your AI policy then becomes rules the system applies rather than a document people are asked to read.
When Enterprise is still the right call
| Situation | Why Enterprise |
|---|---|
| More than 150 people who must share one Claude organisation | One plan, one admin console and shared projects for everyone |
| Health data under HIPAA | Only Enterprise has the HIPAA-ready offering |
| Your security team requires SCIM for every system | Team has no SCIM |
| You need your own retention period inside Claude | Custom retention is Enterprise only |
| Claude activity must feed your compliance tools | The compliance API and audit logs are Enterprise only |
| You need to pay by invoice | Team is card only |
Choosing Enterprise does not remove the need for a control layer if people also use ChatGPT or Copilot, since Enterprise roles stop at Claude. Deployed OS works on top of Enterprise in the same way as on Team. For the cross-tool view, see AI agent governance and what shadow AI is. The same decision for OpenAI is in ChatGPT Enterprise.
Deployed OS is the control layer described here: your systems connected once, access per user group, company skills and guardrails, and a log of every call, in Claude, ChatGPT and Copilot. It is built by an EU company and you choose where the data is stored. See how the connections work, or join the waitlist below.
Frequently asked questions
How much does Claude Enterprise cost?
According to Anthropic’s pricing page on 3 November 2026, self-serve Claude Enterprise costs $20 per seat per month, billed annually, with a minimum of 20 seats, plus usage charged at API rates. A sales-assisted Enterprise plan with invoicing has no listed price. The total depends on how much people use Claude, since the seat buys access and the work is metered.
What features does Claude Enterprise have that Team does not?
Enterprise adds role-based access with fine-grained permissions, SCIM, audit logs, a compliance API, custom data retention, network-level access control and IP allowlisting, a HIPAA-ready offering, domain capture, Claude Security (beta) and inference hooks. Team already has single sign-on, domain verification, admin controls for connectors, usage analytics, spend controls and org-wide skills.
What is the alternative to Claude Enterprise?
Many companies stay on Claude Team, where usage is included in the seat, and put control over company systems in a layer outside Claude. Anthropic allows a company to buy more than one Team plan, so some run one for the tech team and one for everyone else. An MCP gateway such as Deployed OS then adds access per user group, company skills, guardrails and one log, across every plan and across ChatGPT and Copilot.
Can one company have more than one Claude Team plan?
Yes. Anthropic’s account management FAQ says you can create multiple teams by purchasing multiple Team plans. Each Team plan holds 2 to 150 people and has its own members, billing and admin settings.
When do we need Claude Enterprise?
Enterprise is the right call if more than 150 people must share one Claude organisation, if you handle health data under HIPAA, if your security team requires SCIM, if you need custom retention inside Claude, or if Claude activity must feed your compliance tools through the compliance API. It is also the route if you need to pay by invoice, since Team is card only.
Found this useful? Send it to someone who needs it.