OpenAI launches ChatGPT Work to challenge workplace software rivals

OpenAI used its July 9 rollout to push ChatGPT deeper into office software, introducing ChatGPT Work alongside GPT-5.6 and framing both as tools for white-collar tasks that can live inside one interface. The new agent is built to handle longer, more involved work, including research and analysis across connected apps and files, and to produce finished documents, spreadsheets, presentations, reports and Sites. OpenAI said ChatGPT Work was rolling out on web and mobile to paid plans, excluding Free and Go.

GPT-5.6 arrived as a three-model family with Sol as the flagship, Terra as the lower-cost option and Luna as the fastest, most cost-efficient version. OpenAI said GPT-5.6 is now available across ChatGPT, Codex and the OpenAI API, with the global rollout moving toward full availability over 24 hours. The company described Sol as a step up in coding, knowledge work, cybersecurity and science while improving performance per dollar, a pricing message that signals how seriously OpenAI is competing on cost as well as capability.

The launch goes beyond a new model release. OpenAI is betting that office workers want one system that can draft, analyze, code and automate routine jobs without jumping between specialist tools, a move that could raise switching costs for companies that build workflows around its products. If ChatGPT Work becomes the place where staff write proposals, assemble presentations and run recurring tasks, OpenAI gains a stronger position inside enterprise budgets and a more direct challenge to the software vendors that already own much of the daily work interface.

Anthropic is the clearest rival in that contest. Its public product lineup includes Claude Code, Claude for Slack, Claude for Microsoft 365, Claude for Chrome and Claude Cowork, all aimed at the same workplace use cases. Anthropic has also said Claude Fable 5 can work for days at a time in agent harnesses and is aimed at ambitious coding projects, while its Economic Index report in June examined how people use Claude and how they view AI’s effect on work. The comparison now looks less like a race to build the smartest chatbot and more like a struggle to control the software layer where knowledge workers spend their day.

The timing added another layer of pressure. The GPT-5.6 launch had been delayed last month after the U.S. government requested a pause over security concerns, underscoring how frontier AI releases are now shaped by both regulation and competition. OpenAI’s latest push suggests it wants ChatGPT to be more than a product people open for quick answers; it wants it to be the operating system for routine work.