OpenAI has moved a key part of the custom GPT transition from future roadmap to present-tense product policy. Personal ChatGPT accounts—including Free, Go, Plus and Pro—can no longer create new GPTs or publish new ones to the GPT Store. The restriction is already active, according to an OpenAI Help Center page updated on September 29.
That distinction matters. Earlier messaging around the retirement of custom GPTs pointed toward a December 11, 2026 transition and a migration path to Plugins. The latest support guidance does not merely describe that future change: it says that new GPT creation is unavailable now for personal accounts, and that new GPT Store publishing is unavailable now as well.
What personal users can and cannot do
OpenAI’s wording is direct. Personal ChatGPT accounts cannot create new GPTs. They also cannot publish new GPTs to the GPT Store. The company describes both limits as account-eligibility restrictions rather than technical errors, which means there is no builder-profile step, configuration change or appeal that restores the capability for a personal account.
Existing GPTs are not disappearing today. OpenAI says they remain usable and can still be edited when the existing plan and permission requirements are met. In practical terms, creators who already have a working custom GPT can maintain it for now, but they cannot begin a new one or take a new GPT public through the Store using a personal account.
Managed workspaces are different. In Enterprise, Edu and other managed environments, the availability of GPT features depends on workspace settings and user permissions. A blocked creator in a managed workspace may therefore be facing an admin policy or role restriction rather than the personal-account rule. But for a normal Plus or Pro user, the answer is no longer ambiguous: the new-GPT path is closed.
The timing changes the migration conversation
OpenAI’s support page still frames the broader retirement as a planned transition, with affected Enterprise workspaces scheduled for December 11, 2026 and other plans expected to follow the same timeline. It recommends moving workflows to Plugins, which combine reusable instructions with connected apps and task-relevant information. A migration experience is planned, although its availability may vary by account and workspace.
The new restriction changes the incentive structure. Until now, a creator could reasonably decide to keep experimenting with new GPT ideas while preparing a later migration. That option is gone on personal accounts. Existing GPTs are now legacy assets to preserve, audit and translate; Plugins are the place where new reusable workflows have to begin.
For SEO, content and operational teams, this is not just a product-label change. Custom GPTs often bundled a prompt, uploaded knowledge and an API action into one shareable object. Plugins push teams to think more deliberately about instructions, connected tools, permissions and reusable workflows. That can be more structured and more capable, but it also means the old configuration cannot simply be treated as a permanent distribution channel.
What creators should do now
First, inventory the GPTs that already matter. Export or copy their instructions, record their knowledge-file sources, list their actions and note every external dependency. The migration guidance makes clear that not everything carries over automatically; a workflow should be tested in its replacement form before the original is retired.
Second, separate the durable part of each GPT from the interface around it. Editorial standards, research rules, brand voice and approval steps can become Plugin instructions. APIs and connected applications may need to be rebuilt or exposed through MCP. Knowledge should be cleaned up so that it remains useful when it is no longer embedded in the old GPT configuration.
Third, stop planning new personal-account GPT launches as if the window were still open. The immediate question is no longer “Should we create one more GPT before December?” It is “Which existing GPT workflows deserve to become Plugins first?”
OpenAI’s official Troubleshooting GPTs page is the source for the current restriction. The same page confirms that existing GPTs remain usable and editable under applicable requirements, while Plugins are the recommended destination for the workflows creators want to keep.