New ChatGPT features in October 2026: locked chats, Slides, Pro agents

Locked chats: hide a conversation behind your PIN

OpenAI says locked chats keep selected conversations in your account while hiding them from regular chat history, pinned chats and conversation search. To lock a chat, open its more options menu (•••) and choose Lock chat. Locked chats then live under Projects > Locked chats and open with your four-digit account PIN, or with Face ID, Touch ID or your device unlock method on supported phones.

The release notes describe a gradual rollout to Free, Go, Plus and Pro on the web, iOS, Android and the new ChatGPT desktop app for macOS and Windows; the help page also lists Business, Enterprise and Edu. Codex conversations and local Work conversations cannot be locked, and the feature is not available on ChatGPT for Healthcare.

Sources: OpenAI Help Center: ChatGPT release notes · OpenAI Help Center: Using locked chats in ChatGPT

What locking does not do, and the PIN warning

According to OpenAI's help page, locking does not create end-to-end encryption or a separate encrypted vault, does not change how long OpenAI retains the chat, does not change your Improve the model for everyone setting and does not remove information ChatGPT remembered before you locked it. Locking a chat that already has a shared link does not delete that link.

The most important detail: resetting your account PIN permanently deletes all locked chats, remembered information stored only within them and files protected only by the old PIN. OpenAI says this cannot be undone and Support cannot recover the content. Choose a PIN you will remember before you lock anything you need to keep.

Sources: OpenAI Help Center: Using locked chats in ChatGPT

When locking a chat is actually useful

Locking is a visibility control. It helps on shared screens and family devices: planning a surprise, keeping a health or money question out of the visible list, or presenting your screen without exposing personal threads. It is not the right tool if you do not want a conversation kept at all; in that case delete it and review your data controls.

Before sharing sensitive details with any assistant, ask whether the details need to be in the chat. Removing names and account numbers from a question usually costs nothing and protects you better than any setting.

ChatGPT Slides: build and edit presentations together

ChatGPT Slides creates presentations from a conversation, files or an outline, and you can share an example deck to guide the style and layout. You can edit text, images, shapes and charts directly or ask ChatGPT to make changes, and teammates can edit the same deck in real time and leave shared comments in Space.

You can present directly from ChatGPT. On the web, decks export to PowerPoint, PDF or Google Slides; in the desktop app you can open them in PowerPoint or Keynote when those apps are installed. Slides is rolling out to eligible Pro, Business and Enterprise users with Space access, starting from a request to ChatGPT in Work to create a presentation in Space.

Sources: OpenAI Help Center: ChatGPT release notes

GPT-6 Pro: parallel agents, faster answers and better tools

In the chat tab, GPT-6 Pro now gives faster responses, makes better use of plugins and tools, and can run multiple agents on different parts of a task in parallel. You can follow their progress in the conversation and open completed tasks to see what each agent did.

The update is rolling out to eligible Pro users. In managed workspaces, availability depends on workspace settings and rollout eligibility. The release note does not give speed figures or say how parallel work counts toward usage limits.

Sources: OpenAI Help Center: ChatGPT release notes

Which update reaches which plan?

Reading the notes together, locked chats are the update most people will see first, because they cover every plan from Free to Enterprise. Slides and the GPT-6 Pro agents are aimed at heavier, paid use, and Slides additionally needs Space access. Gradual rollouts mean two accounts on the same plan can see different things for a while.

If you are deciding whether to upgrade for these features, wait until they appear for an account you can test. A demo of parallel agents does not tell you whether your own tasks split well into independent parts.

• Free, Go and Plus: locked chats, as the rollout reaches your account.
• Pro: locked chats, plus the GPT-6 Pro update and Slides where you have Space access.
• Business, Enterprise and Edu: locked chats; Slides with Space access on Business and Enterprise; GPT-6 Pro availability follows workspace settings.

Write a task that splits well into parallel parts

Parallel agents help most when a task has parts that do not depend on each other: researching three options, checking several documents, or drafting sections that are merged later. They help least when each step needs the previous answer. Before you start, name the parts, the output you want from each and how the results should be combined.

The brief below is our suggestion, not an OpenAI template. It also works with a single model in any chat assistant: the structure makes the answer easier to check even without parallel agents. Replace the bracketed details first.

I need to decide [decision] by [date]. Split the work into three independent parts: 1) [option or question A], 2) [option or question B], 3) [option or question C]. For each part, give the key facts, the cost or effort, and anything you could not verify. Then merge the results into one comparison table, flag where the parts disagree, and finish with a recommendation and the two questions I should answer before deciding.