OpenAI has received approval from the Trump administration to publicly release GPT-5.6, its latest AI model, and simultaneously unveiled a new AI agent called ChatGPT Work. The rollout began in July 2026, roughly two weeks after GPT-5.6 had been restricted to government-approved organizations during a “limited preview” period due to regulatory scrutiny.
CEO Sam Altman described GPT-5.6 as “the best model we have ever produced.” The model suite consists of three variants — Sol, Terra, and Luna — with Sol positioned as the most powerful, targeting coding, cybersecurity, science, and computer use capabilities. OpenAI is also marketing GPT-5.6 as a lower-cost alternative to competitors’ most powerful models, amid industry-wide concerns about rising AI lab costs being passed on to customers.
ChatGPT Work, launched alongside GPT-5.6, is described as a combination of ChatGPT and Codex, designed to bring Codex’s capabilities to non-technical users for non-coding tasks. According to OpenAI, it can “gather context from the apps, files, and workflows you choose and create finished materials such as documents, spreadsheets, presentations, and web apps.” A unified plugins directory connects the tool to services including Slack, Gmail, Google Drive, calendars, and CRMs.
Access is rolling out in stages. Mac and Windows desktop app users worldwide — including free ChatGPT users — have immediate access. On mobile and the web, Pro, Enterprise, and Edu users get access first, with Plus and Business users to follow “over the next few days.” OpenAI said the global rollout will reach full availability within 24 hours of launch.
ChatGPT Work is a direct competitor to Anthropic’s Claude Cowork, which combines Claude and Claude Code. OpenAI and other companies, including Anthropic, Google, and Apple, have been competing to develop AI agents useful to everyday consumers, a goal the source notes has so far produced “varying results.”
Source: The Verge