Personas work in Claude Code, Cursor and Codex, too. See how
  1. Home
  2. Comparisons
  3. Custom GPTs

Custom GPTs are retiring. Keep your know-how.

OpenAI plans to retire custom GPTs on 11 December 2026 and moves them to plugins and skills in ChatGPT. A good moment to keep your team's know-how where it works in every AI tool.

What changes on 11 December.

According to OpenAI, custom GPTs stop running on 11 December 2026. OpenAI says the dates can still change, and Enterprise workspaces with an approved deferral can use them until 11 February 2027.

  • OpenAI's migration turns a GPT's instructions into a skill and its knowledge files into reference files of a plugin
  • Custom actions and the chosen model are not carried over, and a migrated plugin starts private
  • Plugins work in ChatGPT and Codex
  • If your team also works in Claude, Gemini or Cursor, you can keep the same know-how once in sqemes

Move it in three steps.

From a custom GPT to a playbook your whole team can use.

Step 1

Turn instructions into a skill

Copy your GPT's instructions into a sqemes skill and attach its knowledge files as context files. Recurring tasks become prompts with fields to fill in.

Step 2

Decide who sees it

Set who may see and who may edit each playbook. Every saved change keeps a version you can restore.

Step 3

Use it in ChatGPT and beyond

The browser extension brings it into ChatGPT, Claude and Gemini. On ChatGPT Business and Enterprise, an admin can also connect sqemes as an MCP server.

Custom GPT and sqemes.

FeaturesqemesCustom GPT
What it isA library of skills, prompts and personas for your teamA version of ChatGPT set up for one purpose
Where it worksIn ChatGPT, Claude and Gemini through the browser extension, any MCP client, sqemes ChatIn ChatGPT
FilesContext files, stored once and attached to any playbookKnowledge files per GPT
SharingPer playbook: everyone, named people and groups, or only me; who may edit is set separatelyDepending on the workspace: specific people, the workspace, a link or the GPT Store
StatusAvailableRetirement planned for 11 December 2026

As of: 14 November 2026, from OpenAI's help centre. Sources: OpenAI: GPTs in ChatGPT, OpenAI: custom GPT retirement FAQ, OpenAI: plugins, OpenAI: custom MCP servers.

Plugin or sqemes?

OpenAI's own path works well if everyone stays in ChatGPT. You can also do both.

ChatGPT plugins fit when

your whole team works in ChatGPT.

  • You use only ChatGPT and Codex
  • You want OpenAI's migration to move your GPTs for you
  • Your admins manage plugins in the ChatGPT workspace

sqemes fits when

your know-how should not depend on one tool.

  • Your team also uses Claude, Gemini or Cursor
  • Personas should load the right playbook by themselves
  • Every change needs a version you can restore
  • You want to run it on your own servers

More comparisons.

Questions, answered.

What teams ask about moving from custom GPTs.

Still unsure?

Write to us. A real person answers.

Contact us
01When are custom GPTs retired?
According to OpenAI, on 11 December 2026. OpenAI says the dates can change, and Enterprise workspaces with an approved deferral can use them until 11 February 2027.
02Can we use OpenAI's migration and sqemes?
Yes. You can move a GPT to a plugin in ChatGPT and keep the same know-how in sqemes for every other tool. Through the extension or MCP, ChatGPT uses your sqemes playbooks too.
03Does sqemes import a GPT automatically?
No. You copy the instructions into a skill and upload the knowledge files. sqemes Chat, or an AI tool connected over MCP, can help you turn the instructions into a skill.
04What happens to our files?
In sqemes, files are stored once as context files and attached to the playbooks that need them. A change to a file reaches every playbook that uses it.
05Who wrote this page?
sqemes. Every statement about custom GPTs comes from OpenAI's help centre, linked below the table, as of November 2026. If something has changed, tell us and we correct it.