OpenAI custom GPT creation is no longer available on personal ChatGPT accounts, cutting off Free, Go, Plus and Pro subscribers from building or publishing new customized GPTs. The change leaves existing GPTs intact but shifts future creation toward managed Business, Enterprise and Edu workspaces, according to OpenAI’s current support documentation.
The restriction marks a significant reversal for a feature that OpenAI originally presented as a way for ordinary users to build specialized versions of ChatGPT without writing code. When GPTs launched in November 2023, OpenAI promoted the concept as an accessible tool for individuals, educators, professionals and businesses to package instructions, knowledge and capabilities into reusable AI assistants through its original GPT launch announcement.
OpenAI Custom GPT Creation Moves to Workspaces
OpenAI’s current GPT documentation says new GPT creation and publishing are unavailable on personal ChatGPT accounts, explicitly naming Free, Go, Plus and Pro. Users can still interact with GPTs to which they have access, while existing personal GPTs remain usable and may continue to be edited when the user’s subscription and permissions allow it.
The company now directs new GPT creation toward Business, Enterprise and Edu workspaces, where the ability to create, edit, share or publish GPTs can be controlled by administrators and role permissions. That means a feature once positioned prominently for individual creators has effectively become part of OpenAI’s managed-workspace offering for anyone who wants to build something new.
The distinction matters for paid subscribers. A Plus or Pro customer is still paying for a personal ChatGPT subscription, but that payment no longer makes the account eligible to create a new custom GPT because OpenAI now draws the line around account type rather than simply whether the user pays for ChatGPT.
Existing GPTs Are Not Being Deleted
The change does not amount to an immediate shutdown of the existing GPT ecosystem. OpenAI says previously created GPTs remain available for use, and eligible users can still edit existing GPTs rather than being forced to abandon assistants they have already configured under its current sharing and publishing rules.
That is an important limitation on the scope of the change. A user who spent months refining a writing assistant, research tool or workflow-specific GPT does not automatically lose access to it, but that same personal account cannot create a second new GPT to address another task.
Publishing is also restricted for personal accounts. OpenAI’s sharing guidance says Free, Go, Plus and Pro users cannot publish new GPTs, while Business, Enterprise and Edu workspaces retain publishing options subject to organizational settings and permissions.
The practical result is a two-tier system. Personal users can continue consuming much of the existing GPT ecosystem and maintaining eligible creations they already own, while managed workspaces retain the ability to expand that ecosystem with newly created GPTs.
A Sharp Shift From OpenAI’s Original Pitch
The current policy stands in obvious contrast to the language OpenAI used when GPTs debuted. In its November 2023 announcement, the company said users could make customized versions of ChatGPT for themselves, for a company or for everyone, and emphasized that building one required no coding.
OpenAI went further by arguing that many of the most useful GPTs would come from the broader community. It described educators, coaches and other individual builders as potential creators and promoted the GPT Store as a marketplace where public creations could be discovered and, eventually, potentially monetized.
That community-builder model is now substantially narrower. Individual creators who previously viewed Plus or Pro as the natural home for experimenting with specialized AI tools can maintain what they already built, but they cannot start a new GPT unless they move into an eligible managed workspace.
OpenAI has not said in the cited support documentation that existing GPTs are about to disappear altogether. The safer reading is that the company is restricting the creation pipeline for personal accounts while preserving legacy access for now, rather than immediately retiring custom GPTs as a product category.
Projects Offer Some Personal Customization
Personal users still have other methods to structure repeated work inside ChatGPT. OpenAI’s Projects feature is available across free and paid subscription types and lets users group conversations, upload reference files and set project-specific instructions for recurring tasks.
Projects can therefore reproduce some of the practical benefits that made custom GPTs attractive to individual users. A writer can establish recurring instructions and reference materials, for example, while a researcher can keep files, chats and long-running context together in one workspace.
They are not, however, a direct replacement for creating and publishing a standalone GPT. Custom GPTs can contain dedicated instructions, uploaded knowledge, capabilities, apps or externally defined actions, while the GPT publishing system can expose a specialized assistant to other users or to the GPT Store when eligibility rules permit.
For consumers, that distinction means OpenAI still supports personalized workflows but is exerting greater control over which account types can package those workflows into reusable GPT products.
The Business Trade-Off Behind the Change
The policy gives OpenAI a clearer separation between personal subscriptions and managed organizational products. Business, Enterprise and Edu workspaces offer administrative controls over creation and sharing, giving organizations a framework to decide who can build GPTs and how those tools can circulate internally or publicly.
From a governance perspective, that structure may make custom GPTs easier to manage as they become more capable. GPTs can incorporate external apps and API actions, which means customized assistants may connect to outside services and transmit information required to perform those tasks, a risk OpenAI flags in its GPT documentation.
For individual paying customers, however, the change also raises a straightforward product question: how much control should a subscription holder retain over capabilities that were previously marketed as part of ChatGPT’s accessible customization model? Removing new GPT creation from both Plus and Pro personal accounts reduces the freedom of individual builders even if other personalization tools remain available.
The larger strategic signal may be more important than the individual feature loss. OpenAI’s original GPT launch emphasized democratized creation, while its current rules place new GPT building inside managed workspaces, representing a shift from treating custom AI assistants primarily as consumer-created extensions of ChatGPT toward treating them more as governed tools inside organizations.
For users with existing GPTs, there is no immediate need to panic because OpenAI says those assistants remain available. But for Plus and Pro subscribers who expected their personal account to remain a platform for building new specialized assistants, OpenAI has closed a door that it once invited individual users to walk through.
