OpenAI rolls out Pro 500, Astra Ultrafast, and always-on ChatGPT agents

A broader ChatGPT update with a clear direction

OpenAI updated its ChatGPT release notes on October 1, 2026, and the change is more than a simple feature drop. The update ties together a higher-end paid tier, a faster model path, and persistent agents that can keep working between conversations.

For developers and power users, that combination matters because it points to ChatGPT becoming a more capable work surface for long-running tasks, coding workflows, and app-connected automations. Rather than treating each chat as a one-off session, OpenAI is building toward tools that can continue work across time and devices.

Pro 500 and the new Ultrafast path

According to OpenAI, the company introduced Pro 500, a new $500/month plan. It includes the highest included usage of the Pro plans and access to Astra Ultrafast in ChatGPT Work and Codex.

Ultrafast is selected from the model picker and uses a faster service tier for GPT-6 Astra. OpenAI says the feature is available only with Pro 500 among the Pro plans, and buying credits on Pro 100 or Pro 200 does not unlock it.

Pro 500 subscriptions are available through ChatGPT on the web at launch. That makes the rollout easy to notice, but the more important detail is how OpenAI is separating ordinary Pro usage from a faster tier meant for heavier workflows.

Codex Cloud extends coding work beyond the session

OpenAI also added Codex Cloud, which lets users start and continue coding tasks from desktop, web, or mobile. The system uses reusable environments with a project’s repositories, tools, and dependencies, and each task gets its own isolated workspace.

That setup is designed for continuity. OpenAI says Codex Cloud tasks can continue while the user’s computer is asleep, which changes how developers can approach longer tasks that do not fit neatly into a single sitting.

OpenAI added Code Review in Codex as well. The feature is meant for reading pull-request summaries and diffs and discussing changes, which makes Codex feel less like a single-purpose assistant and more like a place to follow work as it moves through review.

Always-on dots and cloud-based monitoring

The most notable shift in the update may be OpenAI’s introduction of dots, described as always-on agents in ChatGPT that can keep making progress between conversations. Users can give a dot a goal, connect apps, and define what it can do on its own.

OpenAI says dots are powered by GPT-6 Astra and have their own cloud computer. That matters because it gives the agent a place to keep operating even when the user is not actively in a chat.

The same direction shows up in Codex Security Cloud, which can scan GitHub repositories on demand or on a schedule and keep checking new commits. OpenAI says it runs in the cloud and can continue after the user closes their laptop. Together, these features make ChatGPT feel more like a managed workspace for ongoing tasks than a purely conversational product.

What developers and power users should watch next

The practical takeaway is not that every user needs the highest-priced plan. It is that OpenAI is making a stronger distinction between casual chat and durable work. Pro 500, Ultrafast, Codex Cloud, dots, and Security Cloud all point in the same direction: tasks that outlive a single prompt.

For developers, the useful question is where persistent work helps most. That may be in coding flows that need repeated context, in repository review, or in automations that should keep moving while a device is offline. For teams, the key is how much of that work can be delegated safely to an agent with app access and a defined scope.

What to watch next is whether OpenAI keeps expanding these cloud-based workflows and how broadly the company makes the faster and persistent options available. The October 1 update suggests ChatGPT is being shaped less as a chat window and more as a platform for ongoing execution.