OpenAI clears GPT-5.6 for public release and launches ChatGPT Work for everyday users

OpenAI has begun a worldwide rollout of GPT-5.6 after receiving federal approval to move the model beyond its limited government-only preview, and it is pairing the launch with a new product called ChatGPT Work. The company says the new agent is designed to help ordinary users turn ChatGPT into a more practical workplace assistant for documents, spreadsheets, presentations and web apps.

The timing matters because OpenAI is now trying to convert model performance into a broader business workflow product at the same time it pushes a more capable and lower-cost flagship model into public use. The release also intensifies competition with Anthropic, Google and other AI companies racing to make “agent” software useful enough for everyday work rather than just demos.

OpenAI CEO Sam Altman called GPT-5.6 the best model his company has produced so far. The firm is also pitching the new system as a step forward in coding, cybersecurity, science and computer-use tasks, while emphasizing that it can be cheaper to run than rival top-tier models.

What OpenAI launched, and why it matters

OpenAI is introducing two closely linked products at once: GPT-5.6, its latest model suite, and ChatGPT Work, a new agent layer built on top of that model family. Together, they are meant to give users not only stronger AI responses, but also a more useful way to turn prompts into finished work.

That combination is important because the AI market has been moving away from simple chat interfaces and toward software that can take actions inside apps. OpenAI’s move signals that the company wants ChatGPT to become a daily productivity hub, not just a conversational tool.

From restricted preview to public rollout

GPT-5.6 had been held in a limited preview period and was available only to government-approved organizations while regulatory questions were being sorted out. OpenAI says that phase is now over, and the model is being released more broadly after what it describes as a federal green light from the Trump administration.

The company did not present the rollout as a soft launch for a small group of customers. Instead, it framed the move as a global expansion that will continue over roughly a day until availability reaches the full user base.

OpenAI said the model now has approval to move out of its restricted preview and into public availability, and Altman described it as the company’s strongest release yet.

How ChatGPT Work is supposed to function

ChatGPT Work is designed to combine the familiar ChatGPT experience with capabilities associated with Codex, OpenAI’s coding-focused system, so that non-technical users can delegate more complex tasks. The goal is to let people create polished work products without needing to manually stitch together context from multiple apps and files.

According to OpenAI, the product can collect relevant information from the tools a user chooses, then produce output such as documents, spreadsheets, presentations and even web applications. In practical terms, OpenAI is promising an assistant that can move from chat to task completion with far less hand-holding than earlier products.

Which apps can it connect to?

OpenAI says ChatGPT Work will use a unified plugins directory to connect with widely used workplace tools. The initial list includes Slack, Gmail, Google Drive, calendars and customer relationship management systems, giving the agent access to common communication and file-management workflows.

This kind of integration is central to the product’s pitch. AI agents become more useful when they can see the right context and use it to generate work that feels ready to send, present or share.

• Slack for team communication and message context
• Gmail for email drafting and email-based work
• Google Drive for document and file access
• Calendars for scheduling context
• CRMs for customer and sales workflows

Who gets access first?

Access is being staggered by device and subscription tier. OpenAI says Mac and Windows users around the world, including free ChatGPT users, should see ChatGPT Work and GPT-5.6 in the desktop app right away.

On mobile and the web, the first wave is aimed at Pro, Enterprise and Edu customers. Plus and Business subscribers are set to receive access over the next few days, and the company says the broader rollout should continue globally toward full availability within 24 hours.

Why OpenAI is pushing agents now

OpenAI’s timing reflects the broader direction of the AI industry. The largest model developers have spent the past year trying to move beyond chatbots and into agent software that can complete useful work across desktop and cloud applications.

The challenge is straightforward: it is easy to show an AI model answering questions, but much harder to make it reliable enough to manage real tasks across email, documents, calendars and internal tools. That gap has become one of the defining product races in AI.

How does OpenAI compare with Anthropic, Google and Apple?

OpenAI is not alone in this push. Anthropic has been working on a similar concept with Claude Cowork, while Google and even Apple have also recently signaled interest in making AI systems more helpful for everyday users. Each company is trying to prove that its assistant can be more than a browser tab with a chat box.

The pressure has increased as open-source agent projects, including the viral OpenClaw, have shown how quickly the idea can spread across the industry. But broad enthusiasm has not yet translated into a universally trusted consumer agent, which means the market remains up for grabs.

OpenAI is effectively betting that a tighter connection between its best model and real workplace tools will help it get ahead in the emerging AI agent race.

What is GPT-5.6 trying to improve?

OpenAI says GPT-5.6 is meant to raise the bar on intelligence and efficiency, with special emphasis on coding, cybersecurity, science and computer-use functions. Those are areas where model quality often determines whether the product is merely impressive or actually dependable.

The company is especially highlighting Sol, the strongest member of the suite, as the engine behind that push. Sol is being positioned as the model that gives GPT-5.6 its biggest leap in capability.

Why price is part of the pitch

OpenAI is also selling GPT-5.6 as a lower-cost option compared with competitors’ flagship systems. That matters because the AI sector has been facing mounting complaints about expensive subscriptions, heavy inference costs and a broader squeeze on margins.

For customers, that pressure has translated into uncertainty about whether elite AI models can stay widely accessible. For vendors, it has become a race to balance performance improvements with enough cost control to keep products commercially viable.

• Higher-end models are expensive to run at scale
• Customers are increasingly sensitive to subscription pricing
• Companies want more useful tools without runaway compute bills
• Lower-cost high-performance models can win enterprise adoption

What the launch says about the AI market

The release of GPT-5.6 and ChatGPT Work is about more than one product update. It reflects a broader shift in the AI market from experimentation to workflow control, where the winners may be the companies that can connect models to the systems people already use every day.

That shift is especially visible in the workplace. Businesses are not just looking for AI that writes text. They want AI that can draft a report from scattered notes, summarize customer messages, generate a presentation and maybe even build a simple internal web app. OpenAI’s new bundle is explicitly aimed at that use case.

Why the government approval story is notable

The regulatory path matters because it shows that advanced model deployment is increasingly shaped by policy as well as product readiness. A model being limited to government-approved organizations, even temporarily, is a reminder that frontier AI releases can carry scrutiny beyond ordinary consumer launches.

While OpenAI framed the approval as a straightforward clearance to go public, the episode underscores the degree to which model access, safety review and policy sign-off can affect the timing of even major commercial releases.

Timeline: how the rollout unfolded

OpenAI’s release came in stages, with the model first appearing under restrictions before moving to broader availability. The company is now trying to convert that controlled preview into a mass-market launch.

What users should expect next

For most users, the immediate question is not whether GPT-5.6 exists, but when it will show up in their account and how different it feels in practice. OpenAI says the rollout is already underway globally, though access will not arrive all at once across every platform.

The more important test will be whether ChatGPT Work actually reduces friction in daily tasks. If it can reliably gather context from the right sources and produce ready-to-use outputs, it may become one of OpenAI’s most commercially relevant products so far. If not, it risks joining a long list of ambitious AI agent ideas that sounded stronger on paper than they worked in real offices.

What to watch over the coming days

Users and enterprise buyers will likely focus on four things:

1. How well ChatGPT Work connects to real workplace apps
2. Whether GPT-5.6 feels noticeably stronger in coding and computer use
3. How the product performs on mobile and web versus desktop
4. Whether the lower-cost pitch holds up against rival premium models

OpenAI’s launch is a clear signal that the company believes the next phase of AI competition will be won not just by bigger models, but by agents that can reliably turn model intelligence into finished work. The company has now put its latest bet into the public market.

Whether that bet pays off will depend on the same question facing the entire sector: can AI finally become useful enough to handle the everyday tasks that workers and consumers actually want to delegate?