Model: ChatGPT · OpenAI
On 11 September 2026, OpenAI said it is planning to retire custom GPTs across ChatGPT plans and move them to plugins that pair reusable instructions with connected apps. It changes how anyone builds a presence inside ChatGPT — and it is a good moment to make sure your brand keeps its place in the answer.
- On 11 September 2026, OpenAI announced a planned retirement of custom GPTs across ChatGPT plans, with a migration path to plugins.
- Existing GPTs keep working until a retirement date that varies by plan and workspace, so nothing disappears overnight.
- Plugins combine reusable instructions with connected apps — the same idea as a GPT, folded into ChatGPT’s wider plugin layer.
- This is a change to the third-party layer inside ChatGPT, not a new model and not a change to how ChatGPT ranks brands in ordinary answers.
- If you built a GPT, you have migration work to do so your presence carries over; for everyone else, it confirms plugins are the surface to watch.
What OpenAI actually announced
In its ChatGPT release notes dated 11 September 2026, OpenAI said it is “planning to retire custom GPTs across ChatGPT plans” and will provide a migration path to plugins, which “can combine reusable instructions with connected apps.” The key detail for anyone who relies on a GPT is that this is a planned change, not an immediate switch-off: you can keep using existing GPTs until the retirement date that applies to you, subject to your permissions.
OpenAI’s own advice is to prepare now — review the GPTs you depend on, identify who created them and who needs access, and, when migration becomes available, test the migrated plugins and check their sharing settings. Timing and availability will vary by plan and workspace, so the exact dates differ from one account to the next. In short: custom GPTs are being folded into plugins, on a schedule that gives builders time to move.
Why retiring custom GPTs matters for visibility
Custom GPTs were, for many brands, their first foothold inside ChatGPT — a branded assistant a customer could open and use directly. Retiring them and consolidating everything onto plugins reshapes that third-party layer into a single surface. It is the same consolidation we saw when the plugin directory replaced the earlier app directory in ChatGPT Work: OpenAI is steadily making plugins the one place where outside tools and brands plug into ChatGPT.
That consolidation is worth watching for the same reason we treated Sign in with ChatGPT as a platform story rather than a login tweak: the plugin directory is becoming the front door through which more of the ChatGPT journey begins. If your brand had presence as a GPT, the migration decides whether that presence carries over or quietly lapses — and being absent from the surface everything is moving toward is not where you want to be.
What the plugin migration changes
OpenAI frames plugins as the successor to custom GPTs, combining instructions with connected apps.
Custom GPTs and plugins — what changes
A custom GPT was a packaged assistant: a set of instructions, sometimes with files or an action, that a user opened on purpose. A plugin does much of the same job but sits inside ChatGPT’s broader connected-apps framework, pairing reusable instructions with live connections to real tools and data. For builders, the practical shift is from a standalone assistant you published to a plugin you configure, share and maintain within the plugin directory.
For most brands, the honest framing is that neither a GPT nor a plugin is a shortcut to being recommended. Both are tools a user chooses to add — not levers that make ChatGPT name you more often in an ordinary answer. What the migration really changes is the plumbing of the third-party layer, and whether the presence you built survives the move. The way ChatGPT decides who to mention in a normal reply follows a different logic, one we set out in how LLMs decide which brands to cite.
What this does and does not change
It is easy to over-read a headline like this, so it is worth being precise. This is not a new model, and it does not change how ChatGPT reasons or how recent releases like GPT-6 Astra answer a question. It also does not, on its own, change who ChatGPT recommends when someone asks a plain buying or comparison question — that has never depended on whether you own a GPT.
What it does change is the third-party surface: custom GPTs are going, plugins are the successor, and any presence you built as a GPT needs migrating so it is not lost. It is a governance and continuity moment more than a ranking one. Treat it the way you would any platform consolidation — protect what you have, move it cleanly, and keep the bulk of your effort on the public evidence that actually shapes answers.
Where ChatGPT still finds its answers
When ChatGPT answers a category question, it draws on its own knowledge, the open web when it browses or runs Deep Research, and any apps the user has connected. The open web is the part you can shape for a broad audience — it is where the model discovers and describes brands it was not already told about. That does not change because custom GPTs are moving to plugins; if anything, it becomes the more reliable place to invest.
Because the leading models lean on trusted third-party material — the same instinct behind preferred sources in AI Overviews — consistency across your own site and reputable external pages is what earns confident mentions. Thin or manipulative pages, the kind spam policies applied to AI responses are designed to catch, are a poor bet. Substance is the one that keeps paying off, on ChatGPT and on every other model.
What this means for your content
The response to a surface change is not to chase it, but to make your fit unmistakable to any capable model. Spell out who your product is for, the problems it solves, the constraints it suits and the outcomes it delivers, in clean, structured content a model can read and reuse. That is the groundwork we set out in is your brand ready for AI search?, and it is the heart of a durable GEO strategy.
Consistency is what a fast, confident assistant rewards: the more your key facts line up across your site, your profiles and third-party sources, the more readily ChatGPT can name you without hedging. The same work pays off across the field — from a multi-model Perplexity to Claude and a desktop Gemini. And if your category can be reached commercially, our overview of ChatGPT advertising covers where paid AI placement is heading — though the organic groundwork comes first.
How to prepare for the migration
If your brand or team built a custom GPT, a little preparation now protects that presence. Here is a short, practical way to get ready — today is 16 September 2026:
- Inventory your GPTs. List every custom GPT your brand or team relies on, and note what each one is actually used for.
- Identify owners and access. Record who created each GPT and who needs continued access, so nothing is orphaned in the move.
- Watch for your retirement date. Timing varies by plan and workspace, so follow the notices for your own account rather than assuming a single deadline.
- Test the migrated plugins. When migration opens, rebuild or import each GPT as a plugin and confirm the instructions and connected apps behave as expected.
- Review sharing and permissions. Check who can see and use each migrated plugin, and tighten or open access to match your intent.
- Check your open-web presence too. Most of your AI visibility comes from the public web the model cites, so confirm your key facts are clear and consistent there while you are at it.
How reconnAI tracks ChatGPT answers
reconnAI monitors how the leading AI models — ChatGPT, Claude, Gemini, Perplexity, Copilot and Google AI Overview — answer questions across regions, and re-baselines that tracking whenever a platform reshapes how brands connect to it. A retirement of custom GPTs is exactly that kind of moment: it does not change the model, but it consolidates the third-party layer onto plugins, and shifts like that are worth watching so you can see where your brand shows up.
It joins a busy run of ChatGPT shifts we have followed lately, from Deep Research to the desktop app. If you want to see how ChatGPT represents your brand today — and whether your presence survives the plugin migration — get in touch with our team or explore how AI visibility tracking works.
Custom GPT retirement: your questions answered
Are custom GPTs being removed right now? No. OpenAI announced a planned retirement on 11 September 2026, with a migration path to plugins. You can keep using existing GPTs until the retirement date that applies to your plan and workspace, so this is a scheduled change rather than an immediate switch-off.
What are plugins replacing GPTs with? Plugins combine reusable instructions with connected apps — much like a custom GPT, but inside ChatGPT’s wider connected-apps framework. When migration becomes available, you rebuild or import your GPT as a plugin and review its sharing settings.
Does having a GPT or plugin make ChatGPT recommend my brand more? Not directly. A GPT or plugin is a tool a user chooses to add, not a lever that surfaces you in an ordinary answer. How ChatGPT decides who to name in a normal reply depends on the public evidence it draws on, not on whether you own a GPT.
What should I do if my brand never built a GPT? Nothing urgent on the migration itself — but the signal is clear: plugins are the third-party surface to watch. Keep your effort on making your fit clear and consistent on the open web, which is what shapes how ChatGPT answers about your category.
About reconnAI
reconnAI tracks how the major AI models represent topics and sources across ChatGPT, Claude, Gemini, Perplexity, Copilot and Google AI Overview — across multiple regions. We monitor how those models answer and how they change over time, so you can stay ahead of shifts in the AI landscape.
Last updated: 16 September 2026.