OpenAI Launches ChatGPT Work with GPT-5.6 Model

OpenAI Launches ChatGPT Work with GPT-5.6 Model

OpenAI has launched "ChatGPT Work," a new product based on GPT-5.6 and Codex technology, capable of autonomously working on complex projects for several hours.

The application appears to be a step towards merging ChatGPT and Codex into a single "super application." Many features already exist. Similar to Anthropic's transition from Claude Code to Cowork, this is likely a rebranding of the coding tool for knowledge work.

ChatGPT Work is being gradually rolled out on web and mobile for different subscription levels, but it is immediately available to all users via the desktop app. Billing is usage-based and adapts to the complexity of tasks.

With ChatGPT Work, OpenAI aims to transform its chatbot into an autonomous worker. The new product combines Codex technology with third-party integrations and operates on the brand new GPT-5.6.

About two weeks after the Trump administration restricted access to GPT-5.6 for certain organizations, OpenAI has now received approval for public deployment. The company is also introducing a new product called ChatGPT Work for mobile, web, and desktop.

OpenAI describes ChatGPT Work as an agent within ChatGPT capable of operating across applications and files, working for hours on complex projects, and delivering finished results like an Excel or Word document based on a given task.

The product leverages Codex technology, which until now was primarily reserved for developers. The new application likely represents a step towards merging ChatGPT and Codex into a single "super application."

Plugins Are Back

The biggest novelty in ChatGPT Work is a new "Unified Plugin Directory" that consolidates third-party integrations in one place. At launch, the directory includes Google Drive, SharePoint, Slack, Microsoft Teams, Gmail, Outlook, Salesforce, Adobe, Zoom, LinkedIn, GitHub, Canva, and Dropbox, among others. Users can call specific plugins with an "@" mention or let ChatGPT determine which data source is relevant on its own.

This approach echoes OpenAI's unsuccessful attempt at plugins in 2023. Greg Brockman, co-founder of OpenAI, recently admitted that those plugins "didn't work at all because the models weren't ready." With GPT-5.6 behind it, OpenAI expects a different outcome this time.

According to OpenAI, ChatGPT Work can handle entire workflows from a single prompt. For example: transforming customer research into a campaign brief, generating marketing assets from that, and adapting them for different markets, all while maintaining context at each step.

What’s Actually New?

From the announcement, it is difficult to discern what is truly new. Many of the features presented, such as Scheduled Tasks, plugins, or Computer Usage, already exist in ChatGPT or Codex, and ChatGPT could already handle long tasks and connect to data sources previously. Similar to Anthropic's transition from Claude Code to Cowork, this mainly seems like a rebranding of Codex for knowledge work with access to local files, as long as users run the desktop application. Anthropic also recently launched Cowork for web and mobile, likely a preemptive move as they knew ChatGPT Work was coming.

Regarding security, OpenAI highlights an "Auto-Review" feature where advanced models check important actions before execution. OpenAI claims that Auto-Review blocked 100% of attempts to extract protected data during adversarial testing.

Gradual Rollout and Usage-Based Billing

On web and mobile, Pro, Enterprise, and Education users have first access to ChatGPT Work. OpenAI states that Plus and Business customers will follow in the coming days. However, via the ChatGPT desktop app for Mac and Windows, the product is immediately available for all plans, including the free tier. At the time of writing, the Windows link on the download page still leads to the Codex app.

Usage operates differently from standard chat. According to OpenAI, ChatGPT Work shares the same consumption pool based on agents with Codex, ChatGPT for Excel, and Workspace Agents. The consumption of a task depends on its size, complexity, and the selected model. Enterprise and Education administrators can use Spending Controls to manage limits at the workspace, group, and individual user levels.

The Codex pricing page lists included quotas and additional credits, although OpenAI notes that the examples presented are based on coding tasks and that consumption for work tasks may differ. Once the quota is exhausted, some Codex Plus and Pro users may purchase additional credits, while others must upgrade to a higher tier or wait for their quota to reset.