Op-ed · Codito Blog

Claude for business: pricing, privacy and which plan to choose in 2026

Claude Team from $20 per seat per month, Claude Enterprise at $20 plus usage at API rates, 20 seats minimum: the official prices checked, what Anthropic does with your data, and the plan to choose for your company's size.

Emile ChalméEmile ChalméFounder of Codito Ergo Sum, Certified Anthropic Partner
7 min read
ClaudePrixConfidentialité
Contents
  1. What is Claude's price by plan?
  2. How much does Claude cost for your team?
  3. What does Anthropic do with your data?
  4. Does Claude meet the privacy requirements of a French SME?
  5. Which Claude plan should you choose for your company's size?
  6. Who is Claude not the right choice for?
  7. How do you deploy Claude safely in 5 steps?

Claude for business comes in 2 plans: Team, from $20 per seat per month for 2 to 150 seats, and Enterprise, at $20 per seat per month with a 20-seat minimum, plus usage at API rates. On these 2 plans, Anthropic does not use your content by default to train its models. These figures were taken from Anthropic's official pages on 4 November 2026.

Transparency: Codito Ergo Sum is a certified Anthropic partner and deploys Claude for its clients: the subscription is opened in your name and at your expense. This guide is therefore not neutral, but every figure points to its source. We also deploy other models depending on the task (see which AI to use in business). Prices may change at Anthropic's discretion: always check the official page before deciding.

What is Claude's price by plan?

Claude has 6 plans, from $0 to $200 per month for one person and from $20 to $125 per seat for a team, excluding usage on Enterprise. All prices are in dollars and exclude taxes.

PlanListed priceIntended forClaude Code
Free$0Individual discoveryNo
Pro$17 per month on annual billing ($200 paid upfront), $20 monthly1 personYes
Max$100 per month, 5 times the usage of Pro1 person, intensive useYes
Ultra$200 per month, 10 times the weekly usage of Pro1 person, very intensive useYes
Team, Standard seat$20 per seat per month on annual billing, $25 monthlyTeams of 2 to 150 seatsYes
Team, Premium seat$100 per seat per month on annual billing, $125 monthly, 5 times the usage of a Standard seatHeavy users in the teamYes
Enterprise$20 per seat per month on annual billing, 20 seats minimum, plus usage at API ratesLarge organisationsYes

Source: claude.com/pricing, accessed on 4 November 2026. Usage limits apply to all plans; how fast you reach them depends on the model, the length of the exchanges and the size of the files.

How much does Claude cost for your team?

For 10 people on Standard seats, Claude Team costs $200 per month on annual billing, which is $2,400 per year, or $250 per month with no annual commitment. Here are 3 simple calculations based on the official prices, excluding taxes.

  • 5 people, Standard seats, monthly billing: 5 × $25 = $125 per month.
  • 10 people, including 2 intensive users, annual billing: 8 × $20 + 2 × $100 = $360 per month, which is $4,320 per year.
  • 30 people: $600 per month on annual Team Standard. On Enterprise, the seats also cost $600 per month on annual billing, but usage is added at API rates.

The 3rd example shows the real trade-off: at equal seat price, Team includes a usage allowance shared between chat, Cowork and Claude Code, while Enterprise bills consumption. Older Enterprise contracts may follow a different seat model: check yours. To control spending, Enterprise lets you set limits per user and per organisation.

What does Anthropic do with your data?

On commercial plans, Anthropic does not use your inputs and outputs by default to train its models. The commercial terms prohibit it for customer content and specify that you own the outputs, subject to complying with those terms.

QuestionFree, Pro, Max, UltraTeam, Enterprise, API
Model trainingSetting chosen by the user, to be switched off if you refuseNo by default
Retention for Claude Code5 years if training is accepted, 30 days otherwise30 days as standard
Deleted conversationsRemoved from history immediately, erased from servers within 30 days according to the privacy centre
Content flagged for usage policy violationsInputs and outputs kept for up to 2 years, classification scores for up to 7 years
Custom retentionNoCustom controls on Enterprise

3 exceptions deserve a line in your internal policy:

  • User feedback. A click on the thumbs up or thumbs down sends the whole conversation concerned, kept for up to 5 years. Owners of a Team or Enterprise plan can switch this feedback off with the "Rate chats" setting.
  • Voluntary programmes. An organisation can choose to provide data, for example through the development partner programme. It must have accepted it.
  • Personal accounts. An employee on a personal Pro account falls under the consumer terms, even for a company file.

Does Claude meet the privacy requirements of a French SME?

Claude can suit a French SME provided you check 4 points: the contract, the location, the access rights and the data your team puts into it. The plan alone does not make you compliant.

  1. The contract. The data processing addendum names Anthropic as processor and includes the standard contractual clauses; the list of sub-processors is public.
  2. The location. According to the privacy centre, data is stored in the United States, and traffic may be routed to a selection of countries in the United States, Europe, Asia and Australia. If your client or your sector requires hosting in France, Claude through direct access does not meet that request: put the question to Anthropic, or look at a sovereign solution.
  3. The access rights. Team brings SSO and administration controls. Enterprise adds role-based permissions, SCIM, audit logs, the compliance API and IP address restriction.
  4. The data entered. The CNIL (French data protection authority) recommends framing uses through internal policies or charters, and reminds us that the user organisation is liable in case of misuse. Our AI policy template starts from this principle.

For security attestations, Anthropic's documentation points to its trust centre, which gives access to the SOC 2 Type 2 report and the ISO 27001 certificate. Your security officer can request and read them. For very sensitive data, such as professional secrecy, the CNIL generally favours on-premises deployment: such a case deserves specific scoping before any choice of tool.

Which Claude plan should you choose for your company's size?

Choose Pro on your own, Team from 2 people and Enterprise beyond 150 seats or as soon as a compliance need requires it. This is our recommendation as an integrator, to be weighed against your constraints.

Your situationRecommended planWhyTo check
Freelancer or sole company headPro, or Max if usage is intensiveLow cost, Claude Code includedSwitch off training in the privacy settings
2 to 19 peopleTeam, Standard seatsCommercial terms, central billing, SSO1 or 2 Premium seats only for heavy users
20 to 150 people, no audit requirementTeam, Standard and Premium mixedPredictable budget, at equal seat priceUsage limits: measure them for 1 month before widening
20 to 150 people, regulated sector or demanding clientsEnterpriseSCIM, audit logs, compliance API, custom retentionEstimate usage at API rates with a pilot group
More than 150 peopleEnterpriseTeam is capped at 150 seatsContracting route, self-service or through the sales team

For small businesses, Anthropic also offers Claude for Small Business, ready-to-use workflows for accounting, prospects or proposals. The page states "Pro for one person, Team for a business of more than one person", monthly, and specifies that nothing goes out without your approval by default. Check that the connectors offered cover your French software.

Who is Claude not the right choice for?

Claude is not the first choice if your company requires exclusively European hosting without exception, or if your whole team already lives in Microsoft 365 or Google Workspace with an assistant included. In the first case, a solution such as Mistral AI hosted in France should be looked at first. In the second, adoption sometimes weighs more than model quality: test the 2 on the same real cases. Finally, if you have only a few occasional uses and no sensitive data, a single seat is enough to start.

How do you deploy Claude safely in 5 steps?

A simple Claude deployment takes 5 steps, and the first 3 are done before opening a single extra seat.

  1. List the personal accounts already in use, with no sanctions, to move usage across.
  2. Write a short policy: authorised tools, prohibited data, human review.
  3. Have the data processing addendum and, if needed, the impact assessment validated by your data protection officer or your lawyer.
  4. Open Team for a pilot group of 5 to 10 people and switch off feedback if your policy requires it.
  5. Train on real cases, keep a record of the training, then measure usage and time saved for 1 month.

At Codito Ergo Sum, the first session of the AI Partnership is used for this: setting up the Claude environment, in your name and at your expense. The AI Partnership costs €499 or €799 excl. VAT per month, no commitment, cancel at any time, delivered remotely. If you are still unsure about the plan or about privacy, let's talk for 30 minutes: we look at your data and your constraints, and if Claude is not suitable, we tell you so.

Official sources (checked on 4 November 2026)

Back

What it looks like for our clients

Hidden costs quantified in our latest AI audits, with the full case studies.

See all case studies →

Frequently asked questions

How much does Claude cost for a business?

As of 4 November 2026, on Anthropic's official page: Claude Team costs $20 per Standard seat per month on annual billing ($25 monthly) and $100 per Premium seat ($125 monthly). Claude Enterprise costs $20 per seat per month on annual billing, with a 20-seat minimum, plus usage at API rates. Prices exclude taxes.

Does Anthropic use my company's data to train Claude?

By default, no on commercial plans, including Team, Enterprise and the API. There is one exception: if a user sends feedback with the rating buttons or reports a bug, the conversation concerned may be kept for up to 5 years and used. An administrator can switch this feedback off on Team and Enterprise.

Where is Claude's data stored?

According to Anthropic's privacy centre, data is stored in the United States, and traffic may by default be processed in a selection of countries in the United States, Europe, Asia and Australia. A company that requires exclusively European hosting should check this point with Anthropic or consider another solution.

Which Claude plan for a small business?

Anthropic states "Pro for one person, Team for a business of more than one person". From 2 people, Team brings commercial contracts, centralised administration and no training by default. Plan for Standard seats, and a Premium seat only for intensive use.

Is Claude Enterprise GDPR compliant?

No plan on its own makes a use compliant. Anthropic offers a data processing addendum that names it as processor and includes standard contractual clauses. Your company remains responsible for its processing: legal basis, information, register and, if needed, an impact assessment. Your data protection officer or your lawyer must validate.