ChatGPT Enterprise in 2026: What It Is, What It Includes, and What It Does Not

TL;DR

• ChatGPT Enterprise is OpenAI’s sales-quoted plan for deploying ChatGPT across an organisation with central security controls, administration, and contracted support.
• ChatGPT Enterprise is intended for large or regulated organisations that require identity controls, governance, or contract terms beyond OpenAI’s self-serve plans.
• Enterprise includes SSO, SCIM, an admin console, expanded context, usage analytics, and no training on business data by default.
• ChatGPT Enterprise does not maintain shared organisational context across your tools. Sentra provides a company-brain context layer for ChatGPT Enterprise or another assistant, giving each assistant access to shared organisational context.

What is ChatGPT Enterprise?

ChatGPT Enterprise is OpenAI’s contracted plan for deploying ChatGPT across a large organisation. OpenAI launched it on August 28, 2023, as a plan with enterprise security, privacy, and central administration. OpenAI sells the plan through its sales team and provides a custom quote instead of offering self-checkout.

ChatGPT Enterprise supports deployments in which procurement, security, and IT require contractual terms and central control over employee access. The plan supports organisation-wide deployment while giving administrators authority over identity, permissions, retention, and workspace policies. OpenAI also states that its models do not train on business conversations or data by default, according to its enterprise privacy documentation.

Smaller companies that can manage seats through a standard online subscription will usually compare Enterprise with ChatGPT Business. Enterprise becomes relevant when your deployment requires negotiated terms, formal security review, or controls that extend beyond a self-serve workspace.

What does ChatGPT Enterprise include?

ChatGPT Enterprise gives security administrators administrative data controls and compliance capabilities. OpenAI does not train its models on business data or conversations by default, and customers own their inputs and outputs to the extent permitted by law. OpenAI encrypts data at rest and in transit, supports configurable retention, and maintains SOC 2 Type 2 coverage. Authorized administrators can review conversation and GPT audit logs through the Compliance API, according to OpenAI.

Centralized identity controls support organization-wide deployment. SAML SSO handles authentication, while SCIM automates user and group provisioning across workspaces. The Global Admin Console lets administrators manage members, groups, roles, permissions, models, and general settings in one place. Custom roles provide more precise access control. Delegated Group Managers can manage usage limits or feature permissions without receiving full administrator access. Audit logs record relevant policy and configuration changes.

Enterprise workspaces receive expanded context capacity, higher usage allowances, and access to enterprise model options. Exact context limits vary by model, so buyers should confirm current limits rather than rely on the original launch specification. Administrators can decide which models each role may use. Workspace analytics report adoption and usage patterns. Personal Analytics lets members inspect their own activity and privacy-protected workspace trends without exposing another user’s individual activity. OpenAI documents current model access, usage allowances, and Enterprise feature availability in its ChatGPT Enterprise release notes.

How is ChatGPT Enterprise different from ChatGPT Business?

ChatGPT Business gives smaller organizations a self-serve workspace with centralized billing, user administration, SSO, usage visibility, and spend controls. OpenAI requires at least two users for a ChatGPT Business workspace. OpenAI publishes current billing options and per-user prices on its ChatGPT Business plan page. Business also offers Premium seats for organizations that need higher usage limits.

ChatGPT Enterprise serves buyers that need contracted purchasing and deeper administrative controls. OpenAI directs organisations to Enterprise when they require SCIM provisioning, role-based access controls, or data residency. Enterprise also supports sales-led purchasing requirements such as invoice billing, purchase orders, net payment terms, and business associate agreements. OpenAI asks these buyers to contact sales rather than complete a self-serve checkout.

ChatGPT Business fits an organisation that can manage users directly and pay under published self-serve terms. ChatGPT Enterprise fits an organisation whose identity, legal, procurement, or compliance requirements call for a negotiated contract or additional controls. OpenAI quotes Enterprise pricing through its sales process, so Business pricing does not provide a reliable basis for estimating an Enterprise contract.

How much does ChatGPT Enterprise cost?

OpenAI does not publish a fixed price for ChatGPT Enterprise. You must contact its sales team for a quote based on deployment scale, contract terms, and required controls.

ChatGPT Business publishes per-seat pricing and supports self-service purchase. Enterprise uses negotiated contracts that can account for deployment size, invoicing requirements, security review, and support terms. See the full ChatGPT Enterprise pricing breakdown for current plan details and cost factors.

How does ChatGPT Enterprise compare with Claude Enterprise, Microsoft 365 Copilot, and Gemini for Google Workspace?

Buyers can separate these assistants by where each one finds company data, how much material it can process at once, and how the vendor sells it. ChatGPT Enterprise and Claude Enterprise operate independently of a dominant office suite. Microsoft 365 Copilot and Gemini for Workspace integrate directly with their parent productivity suites.

ChatGPT Enterprise suits companies that want ChatGPT as a general assistant across roles. Claude Enterprise offers a similar standalone model, with private knowledge uploads and collaboration through Projects. Anthropic lists a 500,000-token context window on the default model for Claude Enterprise. OpenAI does not publish one fixed ChatGPT Enterprise context limit across all available models. Large context windows let an assistant inspect more material in one request, but they do not preserve knowledge between unrelated conversations.

Microsoft 365 Copilot fits companies whose work already lives in Microsoft 365. Microsoft Graph lets Copilot retrieve permitted emails, meetings, chats, and documents across that environment. Gemini for Google Workspace can use accessible content within Google Workspace apps. Context limits depend on the Gemini model attached to the Workspace edition (Google Workspace).

Pricing separates the procurement paths. Microsoft and Google publish per-user add-on pricing for eligible productivity plans. OpenAI and Anthropic use sales-quoted Enterprise contracts, so buyers must request terms instead of comparing public list prices.

None of the four provides shared organisational context across every company tool by default. Copilot and Gemini retrieve data within governed ecosystems, while ChatGPT and Claude rely on connected sources, uploaded files, and workspace boundaries. A separate context layer can reconcile facts across connected tools and make current organisational context available to multiple assistants.

What ChatGPT Enterprise does not do

ChatGPT Enterprise does not create shared organisational context across every project and tool. Projects group chats, uploaded files, and instructions into separate workspaces. In Enterprise, chats cannot reference conversations in another project or general ChatGPT, according to the OpenAI Help Center. Project instructions also apply only within their project and override global custom instructions.

Project file limits also constrain the information available within each project. Each Enterprise project accepts up to 40 files, and users can upload 10 at once. A project can use the documents available inside it, but separate projects cannot combine their knowledge. For example, an engineering project cannot draw on a customer support project when answering a question that spans both areas.

Shared projects use a project-only context boundary. ChatGPT automatically enables project-only memory when someone shares a project, and the shared project loses access to each member’s context outside it. ChatGPT therefore starts with the information available inside that shared boundary rather than an organisation-wide record.

A company-brain context layer can supply shared organisational context without replacing ChatGPT Enterprise. Sentra serves as the company brain across your tools and makes shared organisational context available to ChatGPT Enterprise or another assistant.

Feature comparison

FAQ

Which one should you choose?

Choose ChatGPT Business for a self-serve workspace with published terms. Choose ChatGPT Enterprise when your organisation requires negotiated procurement, centralised provisioning, role-based controls, or data residency.

Evaluate the context layer separately. Add a company brain when ChatGPT must use shared knowledge across your tools and receive consistent organisational context across conversations.

ChatGPT Enterprise provides security and governance controls, but it does not supply shared organisational context across every company tool. Treat that context as a separate infrastructure layer.