
ChatGPT for Business: Business, Enterprise or API?
In short: for teams of up to 200 people, ChatGPT Business is usually enough, from €21 per user per month. Enterprise pays off when you need data residency, your own retention rules or SLAs. If ChatGPT should run inside your processes instead of just in the browser, you need the API.
What ChatGPT can do in a company. And what it can't.
ChatGPT writes, summarises, researches, analyses files and writes code. With Business and Enterprise it runs in a shared workspace, connected to Google Workspace, Slack, GitHub or Microsoft 365.
What ChatGPT can't do on its own is know your processes. It doesn't know how your quoting works, what sits in your ERP or who signs off on what. And it hallucinates: it states wrong things so convincingly that they sound true. That's why every result needs a person who checks it.
Whether ChatGPT, Claude or Gemini fits your daily work better is covered in our AI tools comparison. This page answers the next question: you want ChatGPT in your company, what do you need to sort out?
Business or Enterprise: the difference at a glance.
| Business | Enterprise |
|---|---|
| For teams of 2 to 200 people | For large companies |
| From €21 per user per month | Individual price via sales |
| Sign-in with SAML, SSO and MFA | Plus SCIM, domain verification and roles |
| No training on your data by default | No training on your data by default |
| Standard retention | Your own retention policies |
| No choice of storage location | Data residency in Europe (EEA and Switzerland) and other regions |
| Standard support | Around-the-clock support with SLAs |
What ChatGPT Business costs.
OpenAI offers two seat types that you can mix in the same workspace:
- Standard: €21 per month. $20 billed annually, $25 billed monthly. For everyday work.
- Premium: €105 per month. $100 annually, $125 monthly. More usage and no 5-hour limit of the standard seat.
Both cover ChatGPT and Codex. From December 2026 both share one usage limit. Enterprise has no list price, billing is by agreement, by credits or by tokens.
As of November 2026, source: OpenAI business pricing (opens in new tab)
Data privacy: what happens to your data.
According to OpenAI, it doesn't train its models on your business data by default, neither on Business nor on Enterprise. Enterprise adds data residency, your own retention policies and encryption with your own keys (EKM).
On Business you can't choose where your data is stored. Data residency in Europe, meaning the EEA and Switzerland, is offered by OpenAI on Enterprise only. Stored there are conversations, prompts, images, uploaded files and custom GPTs. Processing in Europe as well is a separate option for eligible Enterprise customers.
Whether that's enough for you isn't decided by the plan alone. Sort out three things before you start:
- Get the data processing agreement and have your data protection officer review it.
- Clarify where your data has to be stored. If it has to be Europe, you need Enterprise or your own solution via the API.
- Decide what goes into the tool: customer data, HR data, contracts. Without an internal rule, everyone decides for themselves.
Licence or API: when ChatGPT belongs in your processes.
A licence gives every employee a good tool. It doesn't change the process. If enquiries should be pre-sorted automatically, quotes generated from your CRM or documents transferred into your system, you need the API. Then you don't pay per head but per run, and the AI works exactly where the problem sits. That's what we build. We show what that looks like as an AI agency.
What tokens are and what they mean for your bottom line is in EBITTDA and AI costs.
- API prices per 1M tokens, input / output:
- GPT-6 Astra: $10.00 / $50.00, the most capable model
- GPT-6.1 Sol: $2.00 / $10.00, for coding and agents
- GPT-6.1 Terra: $1.00 / $5.00, fast and efficient
- GPT-6 Luna: $0.10 / $0.50, for high volume
- Batch 50% cheaper, data residency 10% surcharge. As of November 2026, source OpenAI
Is a ChatGPT licence enough for you?
For many teams, Business is exactly right. These signals show when it isn't.
Business is enough: You want to get writing, research and analysis done faster.
Business isn't enough: You want work to happen without anyone clicking in a browser.
Business is enough: Everyone works with the tool on their own.
Business isn't enough: Results should land directly in your CRM, ERP or ticket system.
Business is enough: It's enough for you that OpenAI doesn't train on your business data.
Business isn't enough: Your data has to be stored in Europe. Then you need Enterprise or your own solution via the API.
Business is enough: You're under 200 people.
Business isn't enough: You need SLAs, roles and central user management via SCIM.
Rolling out ChatGPT in four steps.
Find use cases
Where does routine work cost you the most time today? Two or three cases are enough to start.
Set the rules
What goes into the tool, who checks results, which plan fits.
Train the team
Nobody learns good prompting and critical checking on the side. That's what our AI training is for.
Measure and expand
After four weeks, look at what really saves time. Where it belongs in the process, the API comes in.
What you usually ask us about ChatGPT at work.
The standard seat costs €21 per user per month, $20 billed annually and $25 billed monthly. The premium seat costs €105 and removes the 5-hour limit. Both can be mixed in the same workspace. Business is for teams of 2 to 200 people, as of November 2026.
According to OpenAI, not by default, neither on Business nor on Enterprise. That applies to the business plans, not to private accounts your employees might use on the side. That's exactly why an internal rule on which data goes into which tool pays off.
When you have more than 200 users, need data stored in a specific region like Europe, require your own retention periods or manage users centrally via SCIM. Enterprise has no list price, you negotiate directly with OpenAI. For most mid-sized companies, Business is enough.
Licences, if your people work with ChatGPT themselves. The API, if ChatGPT should run in the background of your processes, for example sorting enquiries or filling your CRM. Many companies need both: licences for the team, the API for one or two processes.
That isn't decided by the plan alone, but by how you use it. Review the data processing agreement, clarify where your data has to be stored and decide which data goes into the tool. On Business you can't choose the storage location, data residency in Europe is Enterprise only according to OpenAI. When in doubt, your data protection officer decides.
A licence is not an AI strategy.
We have no stake in any particular vendor and we don't sell licences. We look at where AI really saves you time, and then we build it.