OpenAI Launches GPT-5.6 as Top Executive Departs

OpenAI made its GPT-5.6 model family broadly available Thursday after receiving approval from the Trump administration, ending a limited preview period during which the model had been restricted to government-approved organizations, The Verge reported. CEO Sam Altman described GPT-5.6 as "the best model we have ever produced." The family comes in three variants: Sol, the flagship; Terra, an intermediate option; and Luna, a lower-cost tier, according to TechCrunch.

Alongside the models, OpenAI introduced ChatGPT Work, a workplace product that combines ChatGPT with Codex so non-technical users can apply Codex-style capabilities to tasks beyond coding. OpenAI said in a blog post that the tool can pull context from apps, files, and workflows a user selects and produce documents, spreadsheets, presentations, and web apps, with a plugins directory connecting to services such as Slack, Gmail, Google Drive, calendars, and customer-relationship software. Desktop users on Mac and Windows, including free-tier users, get immediate access; on mobile and web, Pro, Enterprise, and Edu accounts come first, with Plus and Business users following over several days as the rollout expands globally over roughly 24 hours, the company said.

On the same day, Fidji Simo, OpenAI's No. 2 executive, told staff she is stepping down from her full-time role, The Wall Street Journal reported. Simo said her medical leave has lasted longer and been harder than expected, and that she will shift to a part-time advisory position.

The context

GPT-5.6 spent about two weeks in a regulatory dispute before its public release. The Trump administration had previously sought to restrict the rollout over concerns about potential misuse of the model's cybersecurity capabilities, according to TechCrunch. OpenAI calls GPT-5.6 its strongest cybersecurity model to date, supporting defensive work such as threat modeling, code review and patching, and blue teaming.

Simo joined OpenAI's board in 2024 and became CEO of Applications in May 2025, a newly created role reporting to Altman that consolidated business and product operations. COO Brad Lightcap, CFO Sarah Friar, and CPO Kevin Weil began reporting to her, while Altman focused on research, compute, and safety. Simo disclosed in April that she was taking medical leave for a relapse of a neuroimmune condition. Before OpenAI, she led Instacart as CEO from 2021, including through its 2023 IPO, and spent more than a decade at Meta, where she ran the Facebook app.

Why it matters

The release puts OpenAI in direct competition with Anthropic, whose Claude Cowork product ChatGPT Work mirrors and whose Fable coding model OpenAI explicitly benchmarked against, claiming Sol scores 80 on the Artificial Analysis Coding Agent Index, 2.8 points above Fable 5, at lower cost. The government approval also resolves, for now, a national-security dispute over the model's cyber capabilities. Simo's departure removes the executive who had consolidated OpenAI's business and product operations, leaving Altman without a clear No. 2 as the company weighs a possible IPO, according to TechCrunch.

What’s next

The GPT-5.6 rollout is expected to reach full availability within about 24 hours of launch, with Plus and Business users gaining ChatGPT Work access on mobile and web over the following days. Altman now faces a search for Simo's successor; no replacement has been named. TechCrunch said it has contacted OpenAI for further information. How ChatGPT Work performs against Anthropic's Claude Cowork remains an open question.