OpenAI is developing a new always-on assistant named 'o' that could perform tasks in the background even when users aren't actively chatting with ChatGPT. The feature briefly appeared in references to a $100 ChatGPT Pro plan before being removed, suggesting the company is in early testing phases for what could be a significant shift in how the AI assistant operates.

What Is This New 'o' Assistant and How Would It Work?

The always-on assistant 'o' represents a departure from ChatGPT's current design, where the AI only engages when users actively open the app or website. Instead, this new tool could maintain background tasks while users are away from their devices or not actively using the interface. Code references discovered online suggest the feature could eventually include email capabilities, with technical indicators pointing to an email suffix configuration alongside the assistant's name.

The feature appeared alongside other premium perks on the $100 ChatGPT Pro plan, grouped with increased Work and Codex usage limits, maximum memory capacity, and 100 gigabytes of file storage. This positioning suggests OpenAI views the always-on assistant as a marquee feature for its highest-tier subscription tier.

How Might an Always-On Assistant Change Your ChatGPT Experience?

• Background Task Automation: The assistant could handle routine tasks like organizing files, summarizing emails, or processing information while you sleep or work on other things, eliminating the need to manually prompt ChatGPT for each action.
• Email Integration: Code references hint at future email handling capabilities, which could allow the assistant to read, summarize, or draft responses to messages without requiring you to copy and paste content into ChatGPT.
• Persistent Memory: Combined with the maximum memory perk on the Pro plan, the assistant could remember context and preferences across sessions, making it feel more like a personal AI colleague than a stateless chatbot.
• Proactive Notifications: An always-on system could potentially alert you when important tasks are complete or when action is needed, rather than waiting for you to check in.

OpenAI has not officially confirmed the 'o' project or provided details about its timeline. The company has kept the feature under wraps, with references appearing and then being removed from public-facing code. A separate internal system called Aeon handles custom workspace agents, but 'o' appears to be a distinct initiative focused on individual user assistance.

When Could This Feature Actually Launch?

While the always-on assistant remains unannounced, timing clues suggest a potential reveal could come soon. OpenAI confirmed that DevDay 2026, its annual developer conference, will take place on September 29 in San Francisco, where the company typically announces new tools and technical capabilities. The conference is expected to feature technical sessions, demos, and new tool announcements, making it a likely venue for unveiling features like the 'o' assistant.

The discovery of the 'o' feature in plan references suggests the feature is beyond the earliest conceptual stages, though it remains unclear whether it will launch as a beta test, a full release, or a future roadmap item. OpenAI's pattern of testing features internally before public announcement means the company could be weeks or months away from a formal reveal.

For ChatGPT users currently on the free or standard plans, the always-on assistant would likely remain exclusive to the $100 Pro tier if it launches, positioning it as a premium differentiator in OpenAI's subscription lineup. The feature's potential to handle email and background tasks could appeal to professionals and power users who rely on ChatGPT for productivity workflows.