--- title: "ChatGPT for Business: What It Costs and Where It Stops" description: "ChatGPT Business and Enterprise: pricing, what you get, and the point a subscription stops being enough." canonical: "https://www.whatsnext-ai.com/blog/chatgpt-for-business" published: "2026-09-29T00:00:00.000Z" updated: "2026-09-30T07:37:14.133Z" --- # ChatGPT for Business: What It Costs and Where It Stops ChatGPT Business and Enterprise: pricing, what you get, and the point a subscription stops being enough. ![Smartphone showing the ChatGPT new chat screen, next to a pair of reading glasses on a patterned desk](https://fgzcpjbyiakhjifaciaj.supabase.co/storage/v1/object/public/media/pexels-bertellifotografia-16094061.jpg) ChatGPT Business costs 26 euro per user per month, or 21 euro on annual billing, from two users up. It adds a shared workspace, admin controls and no training on your company data. Enterprise is quote-only. Most people do not arrive at that question cold. They arrive because they are already on Plus, they keep hitting limits, and someone has asked whether the team should move to a proper business plan. So it is worth starting there. ### What Business adds over a personal plan If the only thing bothering you is running into usage limits, the honest answer is that Business is not primarily a limits upgrade. Standard is not sold as a limits upgrade. What you are actually buying is administration: one workspace instead of a dozen private accounts, central billing, SSO, and a contractual position on your company data. That distinction matters because it changes who should pay for it. If one person is hitting limits, that is a licence question for that person. If your finance lead cannot say how many ChatGPT accounts the company has, or who has been pasting client material into them, that is an organisational problem and Business is the thing that fixes it. The Premium licence is the actual limits answer: five times the Standard usage and no five-hour cap, at a higher per-user rate. You can mix the two in one workspace, which is usually the sensible setup once a handful of heavy users emerge. ### What ChatGPT for business costs OpenAI prices per user per month. These are the figures as shown on OpenAI's European pricing page, inclusive of VAT, read on 28 September 2026: - Business Standard, billed monthly: 26 euro per user per month, as of 28 September 2026 - Business Standard, billed annually: 21 euro per user per month, as of 28 September 2026 - Business Premium: five times the Standard usage with no five-hour cap, at a higher per-user rate - Enterprise: custom pricing through sales Business plans start at two users and are scoped for teams of 2 to 200 people. Above that, OpenAI points you to Enterprise. For a team of ten on annual billing that is roughly 2,500 euro a year. Not a large number. It is, however, one that returns every year, which matters later. ### What teams actually use it for Day to day, the work falls into a few recognisable buckets: drafting content and marketing copy, handling customer service replies, general productivity such as summarising meetings and documents, and thinking work like pressure-testing a plan or interpreting a set of numbers. The step beyond ad-hoc prompting is a custom GPT: your own configured version of ChatGPT with fixed instructions and your own documents attached, shared with colleagues. It is where most teams stop, and for good reason, because it is cheap and it removes the need to re-explain context every time. Beyond that sits the API, which lets developers build ChatGPT into your own software. That sits outside these subscriptions. A custom GPT is genuinely useful. It also still waits for someone to open it. ### What the money buys The difference between a personal plan and a business plan is not the model. It is everything around it: - A shared workspace, so colleagues' work is findable instead of locked in private histories - SAML, SSO and MFA, plus central billing and visibility on usage - No training on your business data by default - Mixed licence types in one workspace The last two do the most work. The training position removes the biggest objection your privacy officer will raise, and mixing licences saves real money past fifteen or so users, since most people do not need the heavy tier. What usually goes wrong at rollout is not technical. Half the team has been on personal accounts for months and quietly stays there, because that is where their history and their habits live. You end up paying for licences alongside the behaviour you meant to replace. ### Your team still has to learn to drive it The difference between companies that get value within a quarter and companies that let the subscription renew unused is rarely the licence tier. It is whether people know how to direct the thing. A team of ten who do not use it still costs 2,500 euro a year. A team of ten who know how to frame a task, supply context and check output get a multiple of that back. Same licence, completely different return. We run [training](https://www.whatsnext-ai.com/ai-training) for exactly this, but the point holds regardless of who does it: budget for adoption, not just for seats. ### Is ChatGPT safe for business use? For most companies, yes, provided you buy the business plan rather than letting people work on personal accounts. That habit is the actual risk, not the product. Business gives you no training on company data by default. OpenAI encrypts data in transit with TLS 1.2 and at rest with AES-256, and access runs through SSO. Under GDPR that is the base you need in order to evidence who can reach what. Data residency is the exception, and it is worth stating plainly. The guarantee that your data stays in a named region is an Enterprise feature at ChatGPT, not a Business one. If you work with data that demonstrably has to stay inside the EU, the entry business plan does not answer that requirement. At that point you are talking to sales about Enterprise, or building something you control. ### Where a subscription stops A subscription gives your team an assistant. For a great deal of work that is enough. For heavy, long-running work the chat window is the wrong environment, and reaching for it anyway is a common mistake. That is what Codex is for, and it is included in Business: agents working in parallel on a project in their own cloud environments, finishing jobs like code reviews, refactors and recurring automations rather than advising on them. Anthropic's equivalent is [Claude Cowork](https://www.whatsnext-ai.com/blog/claude-cowork). We see this go wrong with clients regularly, someone forces a complex job through a chat thread, concludes the model cannot handle it, and the real problem was the environment rather than the model. The fix is to change surface, not to buy a more expensive licence. That moves the boundary a long way out, and it is worth being precise about where it actually sits. Codex can start on its own: on a schedule from hourly to weekly, or on an event in Gmail, Slack or GitHub. Anyone telling you a subscription always waits for a human to press go is wrong. The boundary is somewhere else. Codex is built around code and repositories, and its automations are aimed there too: finding bugs in recent commits, following up pull requests, reporting on your codebase. Your CRM, your quoting system and your document archive are not in that picture. Three moments are where you notice: - The answer has to come from your own documents and systems, not from what the model knows in general - The output has to land in a CRM, a proposal or a case file, rather than in a chat window or a repository - The process runs across systems none of these tools were ever wired into, and someone is still joining the pieces by hand At that point you do not need a bigger subscription. You need something that starts on its own, reads your data and finishes the job. That is [the difference between an assistant and an agent](https://www.whatsnext-ai.com/glossary/chatbot), and it is a difference in kind rather than in size. ### What the alternative costs [More2Win](https://www.whatsnext-ai.com/cases/turning-scattered-knowledge-into-an-ai-powered-research-engine) had years of institutional knowledge spread across more than a thousand documents, and preparing a proposal meant half a day of hunting through folders. We built them a search engine over it. From the published case: - More2Win now answers a question in 30 seconds, where it was half a day of preparation per proposal - More than 1,000 documents are searchable in seconds at More2Win - Per proposal: 320 euro of manual search time, against 50 cents with More2Win's own engine - At 60 proposals a year that is 19,200 euro against 30 euro for More2Win Those 640-times-lower costs per proposal put a debate about 21 versus 26 euro per seat in proportion. Not because the subscription is expensive, but because it answers a different question. Building it was not free either, and that belongs in the picture. The point is what you are left holding. As the case puts it: if What's Next disappeared tomorrow, More2Win still owns the data, the prompts and the configuration. That is the difference between buying a SaaS subscription and building a strategic asset. More2Win also had a hard constraint that the data could not leave the EU and could not train external models. A requirement rather than a preference, and precisely the one an entry-level subscription does not meet. ### Choosing between them A subscription is the right call when the work is different every time and your team mainly needs to write, summarise and think faster. If you already run on Microsoft 365, the same trade-off appears with [Microsoft Copilot](https://www.whatsnext-ai.com/blog/what-is-microsoft-copilot). The difference is mostly where the tool sits. Copilot lives inside Word, Excel, Outlook and Teams and reaches your company files through Microsoft Graph, which makes it strong for work already happening in those files. ChatGPT sits outside your document stack and is freer as a result. On a Microsoft-only estate Copilot usually wins on convenience; on a mixed stack ChatGPT is usually the more practical choice. Both solve the same part of the problem, and both stop at the same place. [Building](https://www.whatsnext-ai.com/ai-agents) becomes interesting the moment a process repeats, the answer has to come out of your own systems, or there are hard requirements about where data lives. In practice most companies do both: a subscription for daily thinking work, and something purpose-built for the two or three processes where the real money sits. We build that in code rather than in a low-code tool, and without tying you to one vendor. We are an [Anthropic](https://www.whatsnext-ai.com/blog/what-is-claude-ai) partner, but we pick the model that fits the task, which keeps the door open when the landscape looks different in a year. Not sure which side your processes fall on? Book a [free advisory call](https://www.whatsnext-ai.com/contact) and we will work out which part a subscription handles well, and which part it does not. Sources: pricing and features as shown on OpenAI's European pricing page, accessed 28 September 2026. More2Win figures are taken from the published case study on this site. ## Frequently asked questions ### How much does ChatGPT for business cost per month? 26 euro per user per month billed monthly, or 21 euro per user per month billed annually, inclusive of VAT and from two users up. Enterprise has no public price. ### Is it worth getting ChatGPT for business? Only if you want the administration. Business adds a shared workspace, SSO, central billing and a position on your company data. If the problem is purely usage limits, the Premium licence is the tier that addresses that. ### Can I use ChatGPT for business on a personal account? Technically yes, sensibly no. A personal account gives you no central administration, no SSO and no visibility into what happens to company data. Those are precisely the reasons to move. ### What is the difference between ChatGPT Business and Enterprise? Business is built for teams of 2 to 200 users with standard administration. Enterprise adds a larger context window, deeper security controls, data residency across ten regions and an SLA, at a custom price. ### Is Codex included in the business plans? Yes. OpenAI lists all features of ChatGPT, ChatGPT Work and Codex with the business plans. Codex is aimed at development work and can start on a schedule or on an event in Gmail, Slack or GitHub.