OpenAI launches GPT-5.6 as ChatGPT takes on your entire workday

For much of the past two years, the race in artificial intelligence has centred on a single question: whose model is smarter?

OpenAI's latest announcement suggests the question is changing.

Instead of simply unveiling another chatbot, OpenAI has introduced a new family of AI models alongside ChatGPT Work, a product designed to complete real workplace tasks by bringing together conversation, coding and automation in a single assistant. The rollout is powered by the new GPT-5.6 family, which includes three models — Sol, Terra and Luna — each aimed at different levels of performance and cost.

The launch marks OpenAI's broad public release of GPT-5.6 after weeks of restricted access. Last month, the company delayed the rollout at the request of the US government while officials carried out additional national security reviews of its frontier AI models. Approval for a wider launch came only after further testing.

But the bigger story may not be the model itself.

It's what OpenAI wants the model to do.

ChatGPT Work is designed to handle jobs that typically require people to move between multiple apps. It can generate documents, analyse spreadsheets, write code, build presentations, and interact with workplace platforms, including Slack, Gmail, Google Drive, and customer relationship management software. Rather than acting as a chatbot waiting for instructions, it is intended to carry out longer, multi-step workflows with limited supervision.

That puts OpenAI on a collision course with rivals pursuing the same goal.

Anthropic recently introduced Claude Cowork, while Google has been expanding Gemini into Workspace. Across the industry, the competition is shifting away from standalone AI assistants toward software that can quietly take over parts of the working day.

OpenAI says GPT-5.6 has also been redesigned around efficiency.

The flagship Sol model targets detailed reasoning and programming tasks, Terra is positioned as the balanced everyday option, while Luna focuses on speed and lower operating costs. The company says the smaller models can deliver much of the flagship's capability while using significantly fewer computing resources, a growing priority as AI companies look to reduce the cost of running increasingly powerful systems.

The release comes at a crucial moment for OpenAI.

The company is pushing deeper into enterprise software just as competition in generative AI intensifies. Meta has begun rolling out its Muse image-generation model across its apps, Google continues to expand Gemini, and Anthropic has focused heavily on AI agents that can complete complex tasks on users' behalf. The race is no longer determined solely by benchmark scores. Increasingly, it's about which company's AI becomes part of everyday work.

For users, GPT-5.6 may look like another model update.

For the industry, it points to something larger.

The era of asking AI for answers is giving way to one in which AI is expected to produce the presentation, update the spreadsheet, review the code, and complete the workflow before the user reaches the next meeting.