Artificial intelligence (AI) company OpenAI has now released the GPT-5.6 frontier model family for everyone, which marks the end of a staggered rollout following a guideline by the U.S. government due to some apprehension about safety guardrails with recent frontier models.
The model family is a troika based on capabilities—Sol, Terra and Luna. This release comes at a time when Anthropic’s Fable model is considered to be setting the capability benchmark, but OpenAI is pitching reliability and versatility, as well as as major advancements in agentic coding, and judgement.
Alongside, OpenAI has also released the ChatGPT Work agentic AI tool, which the company says can “stay with complex projects for hours”.
Among OpenAI’s latest models is Sol, which is the most powerful. The one that it’s in the middle in terms of parameters and size is Luna, with particular focus on speed. Terra, meanwhile, is the most frugal model that will hope to find a balance between the other two for everyday work.
ChatGPT Work finds a basis in Sol’s newfound capability called the ultra mode, which goes beyond the capabilities of a single agent by delegating tasks to various subagents to accelerate complex work. The agentic tool can gather context across connected apps and files to create documents, spreadsheets, presentations and other work. It can also work across the web, phones and computers.
OpenAI cites examples of ChatGPT Work being able to independently turn new messages from Microsoft Teams and Slack into updated docs or slides, and subsequently then share important changes with a group of users.
This is rolling out to the ChatGPT apps for Mac and Windows, for all subscription tiers, with the web expected to roll out soon.
Also Read: Meta challenges OpenAI with its first-ever paid AI API, debuting Muse Spark 1.1
In fact, Anthropic’s recent frontier models, Fable 5 and Mythos 5, were taken offline for a period of time, to comply with an emergency export control directive issued by the U.S. government. The U.S. Commerce Department was alerted that a “jailbreaking” technique could bypass Fable 5’s safeguards, potentially allowing the AI to identify software vulnerabilities.
The government directive specifically prohibited foreign nationals, including Anthropic’s own non-U.S. citizen employees, from accessing the AI models. Since a large portion of Anthropic’s AI lab workforce includes foreign nationals, the company opted to suspend models for all users worldwide to ensure compliance.
OpenAI says almost 1 billion people use ChatGPT every week.
