This website uses cookies

Read our Privacy policy and Terms of use for more information.

Last Updated: November 21, 2026

Claude Enterprise is Anthropic's plan for larger organizations: $20 per seat per month, billed annually, with a 20-seat minimum, plus every token your team uses billed at standard API rates. It adds the controls IT and compliance teams ask for, including SCIM provisioning, audit logs, a Compliance API, custom data retention and a HIPAA-ready option, on top of everything in the Team plan.

The low seat price is the part people misread. On Enterprise, the seat buys access, not usage, so the real bill depends on how hard your people use Claude. Here is what the plan includes, how the billing works, what a realistic year costs, and when Team is the smarter buy. Everything below was checked against Anthropic's own pages on November 20, 2026.

Free download: The AI Stats Sheet 2026. 36 key AI numbers for business on four printable pages, with the source under every figure. Join 2,000+ business leaders who read AI Business Weekly and get it in your inbox right away.

In this guide

What Is Claude Enterprise?

Claude Enterprise is the top business plan for Claude, built for companies that need security, compliance and admin controls across many teams. Anthropic's Enterprise plan page says every seat includes the full set of Claude products:

  • Claude chat on web, desktop and mobile

  • Claude Code in the terminal, IDE, Slack and web

  • Claude Cowork for delegated, multi-step tasks

  • Claude Design for visuals, prototypes and slides

  • Claude for Microsoft 365 (Excel, PowerPoint, Word and Outlook) and Claude in Chrome

Admins decide who gets what through role-based access controls. Claude Enterprise also connects to company tools: Anthropic lists pre-built connectors for Google Drive, Gmail, Slack and Microsoft 365, and custom connectors to internal systems through MCP. Admins approve connectors for the whole organization, and the connectors respect each user's existing permissions.

The plan is not new. Anthropic launched Claude Enterprise on September 4, 2024, with a 500,000-token context window and GitHub integration, according to TechCrunch. Back then Anthropic would not disclose the price. Today the seat price is public, and since February 12, 2026, any company can buy Claude Enterprise self-serve on Anthropic's website without talking to sales. The older "Claude for Work" name is retired.

How Much Does Claude Enterprise Cost?

Claude Enterprise costs US$20 per seat per month, billed annually, with a 20-seat minimum, plus usage billed at API rates, according to Anthropic's pricing page. That makes the smallest possible plan $4,800 a year in seat fees before anyone sends a single prompt.

The usage part is what sets Enterprise apart. Anthropic's billing help article says the seat fee covers access but no usage, and every token your team uses in Claude, Claude Code and Cowork is billed separately at standard API rates. Enterprise seats come with no individual token allowance and no per-seat usage limits. All usage is pooled.

These are the API rates that usage is billed at, per million tokens (MTok), from Anthropic's pricing page:

Model

Input per MTok

Output per MTok

Haiku 5.5 (prompts up to 100K tokens)

$0.10

$0.50

Sonnet 5.5

$2

$10

Opus 5.5

$4

$20

Fable 5.2

$8

$40

Two settings change the rate. Turning on US-only inference bills usage at 1.1x standard rates for Claude Opus 4.6, Sonnet 4.6 and later models, according to Anthropic's billing article. Prompt caching cuts the cost of repeated context: a cache read on Opus 5.5 is $0.20 per million tokens against $4 for fresh input. For the full per-model list, see our Claude API pricing guide.

What a year of Claude Enterprise might cost: a worked example

Here is a hypothetical 100-person company. The usage numbers are assumptions for illustration, not Anthropic figures. Your mix will differ.

  • 80 light users on Sonnet 5.5, each using 1 million input and 200,000 output tokens a month: $2 + $2 = $4 per person a month.

  • 20 heavy users (developers on Claude Code) on Opus 5.5, each using 50 million input and 5 million output tokens a month, with no caching: $200 + $100 = $300 per person a month.

Cost line

Monthly

Yearly

100 seats x $20

$2,000

$24,000

Usage: 80 light users x $4

$320

$3,840

Usage: 20 heavy users x $300

$6,000

$72,000

Total

$8,320

$99,840

That works out to about $83 per seat a month, more than four times the headline seat price. Notice where the money goes: the 20 heavy users drive 72% of the total bill. Light users are cheap. Power users on Opus are not, and caching and model choice matter more than seat count.

Claude Team vs Enterprise: What's the Difference?

Claude Team is a fixed-price plan for 2 to 150 people with usage limits per seat, while Claude Enterprise has no seat cap or per-seat usage limits, bills usage at API rates, and adds deeper security and compliance controls. Here is the side-by-side from Anthropic's pricing and help pages.

Feature

Team

Enterprise

Seat price

Standard $20/month annual ($25 monthly); Premium $100 annual ($125 monthly)

$20/month, annual only

Size

2 to 150 seats

20-seat minimum (self-serve), 50 (sales-assisted)

Usage

Included, with limits per seat (Premium = 5x Standard)

Billed at API rates, no per-seat limits

Claude Code, Cowork, Design, Microsoft 365

Yes

Yes

SSO and enterprise search

Yes

Yes

No training on your content by default

Yes

Yes

SCIM provisioning and audit logs

No

Yes

Compliance API and custom data retention

No

Yes

IP allowlisting and network-level access control

No

Yes

Spend controls

Listed in Anthropic's comparison table

Admin-set limits per user and for the whole organization

HIPAA-ready offering

No

Yes

Pooled API credits

Up to $500/month

Usage is the bill

For the same 100 people in our example, a Team plan with 80 Standard and 20 Premium seats would cost $3,600 a month on annual billing, or $43,200 a year. That is less than half the Enterprise estimate. The trade-off is that Team seats have usage caps, and Team lacks SCIM, audit logs, the Compliance API and custom retention. For many mid-sized companies, those missing controls decide it, not the price. Our Claude pricing guide covers every individual and Team plan in detail.

What Security and Compliance Features Does Claude Enterprise Have?

Claude Enterprise covers identity, visibility, data and network controls, and Anthropic says prompts, data and results are not used to train its models by default. Here is the list from Anthropic's Enterprise page and help center:

  • Identity: SSO (SAML) with domain capture, SCIM and just-in-time provisioning, role-based access and spend controls.

  • Visibility: audit logs covering user actions, system events and data access; a Compliance API with activity logs, chat histories and file content, filterable by user and time; an Analytics API; and OpenTelemetry monitoring.

  • Data: custom data retention, customer-managed encryption keys held in your own cloud provider, and US-only inference.

  • Network: IP allowlisting and network-level controls that can block personal Claude accounts on the corporate network.

  • Compliance: SOC 2, ISO 27001, GDPR and CCPA, plus a HIPAA-ready configuration with a business associate agreement (BAA) that eligible organizations can accept in their settings.

Several of these, including OpenTelemetry and custom data retention, are Enterprise-only, according to Anthropic. If your security team needs to block personal AI accounts or export chat records for legal holds, Team will not do it.

Laptop showing a padlock on its screen, representing enterprise data security

Security controls are the main reason companies pick Enterprise over Team. Photo: Allison Saeng / Unsplash+

Sameer's take: Sameer Khan says the public-sector organizations he works with are actively piloting AI, and their biggest worry is privacy and data security. He would tell a business leader to pay for Claude and ChatGPT, and he warns that the biggest mistake companies make with AI is trusting its output without checking it.

How Do You Buy Claude Enterprise?

You can buy Claude Enterprise three ways: self-serve on Anthropic's website, through Anthropic's sales team, or through AWS Marketplace. The terms differ, according to Anthropic's help center.

Option

Self-serve

Sales-assisted

Minimum seats

20

50

Payment

Credit card or ACH, USD only

Credit card, ACH or invoice; multi-currency

How usage is paid

Prepaid credits that draw down

Billed monthly in arrears

HIPAA and BAA

Available

Available

A few billing details matter in practice. On self-serve, usage stops when credits run out until an owner buys more, and seats cannot be removed mid-term. Seats added mid-year are charged a prorated amount right away. Sales-assisted invoices of $50,000 or more must be paid by bank transfer. Plans bought through AWS Marketplace are billed monthly in arrears through AWS and can draw down an existing AWS commitment. Anthropic also offers tiered incentives on committed spend, trials and customer success support at certain spend levels through sales.

Already on an older Enterprise contract? Anthropic says legacy Chat and Chat + Claude Code seats convert to the single usage-based Enterprise seat at your next renewal, while Standard and Premium seats keep their current limits until you arrange a move with Anthropic.

Who Should Choose Claude Enterprise?

Claude Enterprise fits companies with more than 150 users, regulated data, or security teams that need audit trails and control over where data lives. Team fits smaller companies that want predictable monthly costs.

Business team reviewing data together during a meeting in a modern office

Company size and data rules usually decide between Team and Enterprise. Photo: Getty Images / Unsplash+

Choose Enterprise if any of these is true:

  • You have more than 150 people who need Claude (Team tops out at 150 seats).

  • You handle health data and need a BAA, or need custom retention, encryption keys or US-only processing.

  • Your IT team requires SCIM, audit logs or a Compliance API before approving any tool.

  • You have a group of very heavy users, such as developers on Claude Code, who would keep hitting Team limits.

Stick with Team if you have under 150 users, no special compliance needs, and want a fixed bill. You can upgrade from Team to Enterprise later, according to Anthropic's help center.

Demand is clearly there. In Ramp's AI Index, as reported on May 13, 2026, 34.4% of businesses in Ramp's data paid for Anthropic's services, ahead of OpenAI at 32.3%, up from 9% for Anthropic in May 2025, according to TechCrunch. For more adoption data, see our Claude AI statistics.

How to Keep a Claude Enterprise Bill Under Control

The biggest risk with usage-based billing is a surprise invoice. These steps keep it predictable:

  1. Set spend limits on day one. Admins can cap spend for the whole organization and for each user. The lower limit applies, and usage stops when it is reached.

  2. Default to the cheaper model. Sonnet 5.5 costs half of Opus 5.5 per token. Save Opus and Fable for work that needs them.

  3. Watch your heavy users. In our example, one in five users drove most of the bill. Usage analytics show who they are.

  4. Use caching for repeated context. Long codebases and reference documents read from cache cost a fraction of fresh input.

  5. Run a 60 to 90 day pilot. Measure real token use before committing to more seats or a spend commitment.

The Bottom Line

Claude Enterprise is $20 a seat plus API-rate usage, and the usage is where the money goes. For a light-use workforce it can be cheap. For a team of developers living in Claude Code, plan on several hundred dollars per heavy user each month unless you manage models and caching.

Buy Enterprise for the controls: SCIM, audit logs, the Compliance API, custom retention, HIPAA readiness and no seat cap. If you do not need those, Team is simpler and usually cheaper. Either way, set spend limits first and pilot before you scale.

Related Articles

⭐ Get more AI Business Weekly in Google: add us as a preferred source on Google.

Written by Sameer Khan, with support from the AI Business Weekly Editorial Team. AI-assisted and fact-checked against the sources linked above.