Claude Team Plan for European Teams 2026
Claude Team plan for European teams: Standard 20 or 25 USD, Premium 100 or 125 USD, 2 to 150 seats. DPA yes, no EU inference. Team vs Enterprise and Bedrock.
TL;DR
The Claude Team plan is the first company contract for Claude: Standard seats cost 25 USD a month or 20 USD on annual billing, Premium seats 125 or 100 USD, for 2 to 150 people. You get a DPA, SSO and no training by default. There is no EU inference. Use Bedrock or Google Cloud when processing must stay in the Union.
What does the Claude Team plan cost in 2026?
The Claude Team plan bills per seat, not per token, until someone turns on extra usage credits. Standard is 25 USD a month or 20 USD on annual billing. Premium is 125 USD a month or 100 USD annual. You need two seats and you stop at 150. Prices are US list prices and exclude tax.
The Team help article is the source for those four numbers. The pricing page confirms Team covers 2 to 150 people and that annual billing exists for Team. Anthropic says currency and tax handling vary by region. Visit claude.ai/upgrade for the checkout you actually pay. We did not find a published euro list price on 3 October 2026.
| Seat | Monthly | Annual (per month) | Usage vs Pro | Who it is for |
|---|---|---|---|---|
| Team Standard | 25 USD | 20 USD | 1.25x Pro | Most knowledge workers |
| Team Premium | 125 USD | 100 USD | 6.25x Pro | Heavy Claude Code users |
| Enterprise (single seat) | not offered | 20 USD plus API usage | no included pool | Orgs that need SCIM and audit |
Sources: Team plan, Enterprise plan. The billing article dated 6 August 2026 walks through Anthropic’s own examples, tax excluded:
| Example (Anthropic) | Calculation | Charged |
|---|---|---|
| 10 Standard, annual, 1 January | 20 x 10 x 12 | 2,400 USD on day one |
| Upgrade 5 of those to Premium on 1 April (275 days left) | (80 x 5 x 12) x (275 / 365) | 3,616.44 USD that day |
| 10 Standard, monthly, 1 January | 25 x 10 | 250 USD that month |
| Upgrade 5 to Premium on 15 January (16 days left in a 31-day month) | (100 x 5 x 16) / 31 | 258.06 USD that day |
Adding seats or upgrading Standard to Premium is prorated and charged immediately. Removing a member does not credit the invoice. The empty seat stays on the plan until an Owner reduces the allocation. The “Projected total” on the Billing page excludes tax, so the card charge is higher.
Our own 12-month arithmetic, not a vendor quote: 8 Standard plus 2 Premium on annual billing is (8 x 20 + 2 x 100) x 12 = 4,320 USD before tax. The same mix billed monthly is (8 x 25 + 2 x 125) x 12 = 5,400 USD. Claude Code sits in every Team seat and shares the same usage pool as chat, per the Team and Enterprise Claude Code article. The full Free-to-Enterprise grid is in our Claude pricing plans 2026 overview.
What do Standard and Premium seats actually include?
A Team seat is a workplace product: Claude on web, desktop and mobile, Claude Code, Cowork, Projects you can share, connectors, enterprise search, SSO and central billing. Standard and Premium differ on usage, not on the admin surface. Limits are per person. One colleague at the cap does not stop the others.
The Team article lists SSO, JIT provisioning, spend controls, connectors (Google Drive, Gmail, Calendar, GitHub, Microsoft 365, Slack) and enterprise search across Slack, Microsoft 365 and custom connectors. The pricing comparison table is stricter on identity: Team has SSO and domain verification. Domain capture, SCIM, audit logs, the Compliance API, custom retention, HIPAA-ready configuration and role-based access are Enterprise only. Model training is “none by default” on Team.
| Control | Team | Enterprise |
|---|---|---|
| SSO and domain verification | Yes | Yes |
| Domain capture, account migration | No | Yes |
| JIT provisioning | Yes | Yes |
| SCIM | No | Yes |
| Audit logs, Compliance API | No | Yes |
| Custom data retention | No | Yes, 30-day minimum |
| Spend controls, usage analytics | Yes | Yes |
| Extra usage credits | Yes (optional) | Usage is the product: billed at API rates |
| US-only inference | No | Yes, 1.1x on 4.6 and later |
| EU inference or EU storage | No | No |
Paid plans share a rolling five-hour window and a weekly cap. There is no published message count. Standard is 1.25x Pro per session. Premium is 6.25x Pro, which is 5x Standard. Fable on Team is 50 percent of weekly limits on Premium seats. When a seat is empty, Owners can enable usage credits at standard API rates. Token prices we read on 3 October 2026 from Anthropic pricing:
| Model | Input / output per 1M tokens |
|---|---|
| Claude Sonnet 5.5 | 2 / 10 USD |
| Claude Opus 5.5 | 4 / 20 USD |
| Claude Haiku 4.5 | 1 / 5 USD |
| Claude Fable 5.1 | 10 / 50 USD |
The Team help article still mentions a 200k context window. The pricing table on the same day says “Up to 1M varies by model”. Treat the pricing table as the current product claim and test the model you actually call. Claude Code in the terminal, VS Code, Cursor and JetBrains uses the same seat. Log in with “Claude account with subscription”. An ANTHROPIC_API_KEY in the environment bypasses the seat and bills tokens. Developer-side cost math is in our Claude Code pricing for teams.
Claude Team vs Enterprise: which plan for a European company?
Pick the Claude Team plan when you have 2 to 150 people, you need a company-owned account with SSO, and you do not yet need SCIM, audit logs or a written deletion period. Move to Enterprise when identity offboarding, evidence for an audit or HIPAA-ready configuration is a hard requirement, or when you pass 150 seats.
Enterprise is a different bill. The Enterprise article prices a single seat at 20 USD per month, billed annually, with no usage inside that fee. Self-serve starts at 20 seats (4,800 USD a year before tokens). Sales-assisted starts at 50 seats. Self-serve is USD only on card or ACH. Invoices and other currencies need sales. The seat vs token trade-off against Bedrock is in our Claude Enterprise vs Bedrock EU comparison.
| Situation | Recommendation |
|---|---|
| 2 to about 40 people, company chat and Claude Code | Team, mix Standard and Premium |
| IT must deprovision via SCIM the day someone leaves | Enterprise |
| Works council or DPO wants audit logs and a deletion period | Enterprise |
| More than 150 seats | Enterprise. Team stops there |
| Prompts contain customer or HR personal data that must stay in the EU | Neither Team nor Enterprise. Use Bedrock or Google Cloud |
We recommend Team as the floor for company data. A private Pro or Max account has no DPA, no SSO and belongs to the person. That argument is in our Claude Max for companies piece. Team is not the floor if the legal test is “inference in EU member states”. That test fails on every Anthropic-hosted plan we read today.
Does the Claude Team plan meet GDPR and EU residency?
The Claude Team plan can be a lawful processor setup: Anthropic’s DPA with Standard Contractual Clauses is built into the Commercial Terms. It does not keep inference or storage in the EU. Commercial product data is stored in the US. Traffic may run in the US, Europe, Asia or Australia. There is no EU-only switch on Team.
The DPA help article covers commercial products such as Claude for Work and the API. You accept the DPA (effective 24 February 2025) with the Commercial Terms. Anthropic is the processor, you are the controller. Transfers use SCC Modules Two and Three under Implementing Decision (EU) 2021/914 of 4 June 2021. Third-party Claude (Bedrock, Google Cloud, Foundry) is under that cloud’s terms, not this DPA.
The server location article dated 15 June 2026 says data is stored in the US. The API residency page only offers inference_geo values global and us. Workspace geo is us only. The US-only inference article says the toggle does not appear on Team. On usage-based Enterprise it bills 1.1x for Opus 4.6, Sonnet 4.6 and later. It still does not move storage to the EU.
Retention on Team, from the organisation storage article dated 1 July 2026:
- Saved chats stay so the product works. You delete them in the dashboard. History goes immediately. Back-end storage goes within 30 days.
- Incognito chats are deleted within 30 days unless flagged.
- Usage Policy flags: inputs and outputs up to 2 years, classification scores up to 7 years.
- Feedback (thumbs, bug reports) is kept 5 years.
Custom retention is Enterprise only. The GDPR provider matrix across Anthropic, Bedrock, Google Cloud and Foundry is in our Claude GDPR comparison.
When should a European team pick Bedrock or Google Cloud instead?
Pick Bedrock or Google Cloud instead of the Claude Team plan when the requirement is processing in EU member states, or when you want token billing on an existing AWS or Google commit and you will build your own chat front end. Team wins when people need claude.ai, shared Projects and connectors this week.
Bedrock publishes eu.anthropic.claude-sonnet-5-5 and states that the EU geo ID keeps data within EU regions. In-region on bedrock-runtime is not offered for that model. The sample on the Sonnet 5.5 card still calls global.anthropic.claude-sonnet-5-5 from us-east-1. Copy that sample into Frankfurt and you leave the EU. IAM and SCPs are in our Claude on AWS Bedrock EU guide.
Anthropic pricing applies a 10 percent premium on Bedrock regional endpoints and Google multi-region endpoints from Sonnet 4.5, Haiku 4.5 and Opus 4.5 onward. Google’s partner-model table read on 3 October 2026 lists Sonnet 5.5 at 2.00 / 10.00 USD and at 2.20 / 11.00 USD, and Opus 5.5 at 4.00 / 20.00 and 4.40 / 22.00. That second pair is the 10 percent row.
| Route | What you buy | EU processing | Sonnet 5.5 in / out |
|---|---|---|---|
| Claude Team plan | Seats | No | Included until the seat cap, then API rates |
| Claude API (first-party) | Tokens | global or us only |
2 / 10 USD |
| Amazon Bedrock EU geo | Tokens | Yes, eu. IDs |
10% over global (2.20 / 11) |
| Google Cloud multi-region | Tokens | Europe multi-region listed for current Claude models | 2.20 / 11 USD on the 10% table |
| Microsoft Foundry | Tokens as CCU at 0.01 USD | No EU data zone. Global Standard or US Data Zone | Same list rates, Anthropic is the processor |
Foundry hosting comparison: Hosted on Azure offers Global Standard and US Data Zone. Hosted on Anthropic is Global Standard only. CCU conversion is 0.01 USD per unit, 100 CCU = 1.00 USD of list fees. Per-token EU tables sit in our Claude API pricing EU article.
export CLAUDE_CODE_USE_BEDROCK=1
export AWS_REGION=eu-central-1
export ANTHROPIC_MODEL=eu.anthropic.claude-sonnet-5-5
# Do not copy the us-east-1 global sample from the model card
claude- Confirm the legal test. If prompts must stay in EU member states, stop here and use Bedrock or Google Cloud. Team cannot meet that test.
- Create the Team org at
claude.aiwith a company card. Buy at least two seats. - Turn on SSO and domain verification before you invite anyone. JIT creates accounts on first login. SCIM is not on Team.
- Assign Premium only to people who hit Standard caps. Keep the rest on Standard.
- Set organisation and user spend caps. Enable usage credits only after you have a monthly token budget.
- Point Claude Code at the Team login. Remove
ANTHROPIC_API_KEYfrom developer shells so the seat, not the API, is billed.
FAQ
How much does the Claude Team plan cost per person?
Standard is 25 USD a month or 20 USD on annual billing. Premium is 125 USD a month or 100 USD annual. Those are US list prices and exclude tax. Ten Standard seats on annual billing cost 2,400 USD on day one in Anthropic’s own example. Extra usage after the seat cap is billed at API rates if you enable credits.
Claude Team vs Enterprise: which should a European SME buy?
Buy Team if you have under 150 people and you need SSO plus a DPA this quarter. Buy Enterprise if SCIM, audit logs, custom retention or more than 150 seats are mandatory. Enterprise is 20 USD per seat per year-month plus every token at API rates, with a 20-seat self-serve floor (4,800 USD a year before usage).
Does the Claude Team plan keep data in the EU?
No. Commercial product data is stored in the US. Inference may run in the US, Europe, Asia or Australia. Team has no US-only toggle and no EU toggle. Bedrock eu.anthropic.claude-sonnet-5-5 or Google’s Europe multi-region is the documented EU processing path.
Is there a DPA on the Claude Team plan?
Yes. Team is a commercial product (Claude for Work). The DPA with SCCs is incorporated into the Commercial Terms. Free, Pro and Max are consumer products and sit outside that article. Third-party Claude on Bedrock or Google Cloud uses the cloud contract, not Anthropic’s DPA.
Can we mix Standard and Premium seats?
Yes. Owners buy an allocation of each type and assign people. Upgrades are prorated and charged immediately. A removal does not shrink the bill until you reduce the allocation. Premium is the 6.25x Pro pool, Standard is 1.25x Pro.
What happens when a Team seat hits its usage limit?
The other seats keep working. That person waits for the five-hour or weekly reset, or the org enables usage credits at standard API rates (Sonnet 5.5 is 2 / 10 USD per million tokens globally). Limits reset on a fixed weekday assigned to the account, shown under Settings, Usage.
Sources
- Claude Help Center: What is the Team plan? (3 October 2026)
- Claude Help Center: How is my Team plan bill calculated? (3 October 2026)
- Claude Help Center: What is the Enterprise plan? (3 October 2026)
- Claude: Pricing (3 October 2026)
- Claude Help Center: Enable US-only inference (3 October 2026)
- Claude Help Center: How do I view and sign your DPA? (3 October 2026)
- Anthropic Data Processing Addendum (3 October 2026)
- Anthropic Privacy Center: Where are your servers located? (3 October 2026)
- Anthropic Privacy Center: How long do you store my organization's data? (3 October 2026)
- Anthropic docs: Pricing (3 October 2026)
- Anthropic docs: Data residency (3 October 2026)
- AWS Bedrock: Claude Sonnet 5.5 model card (3 October 2026)
- Google Cloud: Agent Platform pricing (3 October 2026)
- Microsoft Learn: Claude hosting comparison (3 October 2026)