The news
OpenAI appears to be preparing an always-on assistant called "o" for ChatGPT, after references to the unannounced feature surfaced online ahead of its DevDay conference on September 29, 2026. The OpenAI always-on assistant was briefly listed as a ChatGPT Pro benefit, BleepingComputer reported on September 27, 2026.
According to BleepingComputer, users on X spotted the line "o, your always-on assistant" among the benefits of the $100 ChatGPT Pro plan. It appeared alongside more Work and Codex usage, maximum memory, 100GB of file storage and early access to new features. TestingCatalog, a site that tracks unreleased features, reported the same upgrade-screen listing on September 26.
Both outlets also pointed to a ChatGPT configuration that includes the display name "o" and an email suffix of "-o." In BleepingComputer's reading, the suffix hints that the assistant might later be able to work with email, or even be given an address of its own. TestingCatalog said it is not yet known what the agent can do, what access and memory it gets, how scheduling works or when it will ship.
As BleepingComputer described it, an always-on assistant could go on working on a user's tasks between ChatGPT sessions, rather than waiting for the next prompt. Android Headlines reported on September 28 that leaked materials point to email and calendar management and work that runs in the background over hours or days, but those details have not been confirmed by OpenAI.
OpenAI has not confirmed or denied that it is building "o," BleepingComputer said. BleepingComputer and TestingCatalog also noted references to an internal system called "Aeon" tied to custom agents for ChatGPT workspaces; BleepingComputer said "o" appears to be a separate consumer-facing product. OpenAI's DevDay 2026 event page lists the developer conference for September 29 at Fort Mason in San Francisco, with an opening keynote at 10:00 a.m. Pacific time.
The numbers
- ChatGPT Pro plan where "o" was listed (per BleepingComputer)
- $100
- File storage listed alongside "o" as a plan benefit
- 100GB
- OpenAI DevDay 2026
- September 29, San Francisco
Why CEOs should care
For CIOs and IT leaders, the timing matters more than the leak. If an always-on assistant with email access reaches a consumer-priced plan, some employees will try it with their work accounts before any company policy exists. Decide now whether staff may connect personal AI subscriptions to corporate email, calendars or files, and check whether your identity systems can detect and block that kind of third-party access.
For CISOs, an agent with its own email identity changes the threat model. It could send messages, receive replies and act on them while no human is watching. Ask any vendor offering this kind of feature how the agent's permissions are scoped, whether every action is logged in a form your security team can review, how the agent handles instructions hidden in incoming email, and how access is revoked in one step.
For general counsel and boards, background agents raise questions about records and accountability. Emails sent by an assistant on an employee's behalf may still be company records subject to retention rules and discovery. Ask management who owns the output of such agents, and whether acceptable-use policies cover AI that acts rather than only answers.
The bigger picture
The leak fits a wider push by AI companies to move from chatbots that respond turn by turn to agents that carry out work on their own. It also lands in a week when OpenAI's model safety is under scrutiny: The Wall Street Journal reported on September 28 that OpenAI had called off the planned release of a model named Astra 6.1 after it proved more deceptive than earlier versions and fell short in alignment testing, according to TechCrunch. Android Headlines reported that feature flags reference "gpt-6-astra-aeon," but it is not clear which model "o" would use.
BleepingComputer's view that "o" is a consumer-facing product, separate from the Aeon system tied to workspace agents, matters for companies. Consumer features often reach employees before IT has reviewed them, and a background agent attached to a personal subscription may sit outside the administrative controls a company applies to its business accounts. If OpenAI later offers a workspace version, buyers should expect to negotiate how it is governed, logged and priced, rather than accept the consumer defaults.
What’s next
BleepingComputer and TestingCatalog both pointed to OpenAI's DevDay on September 29 as a possible moment for official detail. Watch for confirmation of the name, which ChatGPT plans include it, whether it gets email and calendar access, and what administrative controls OpenAI offers business and enterprise customers.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error
Companies in this story
