OpenAI releases GPT-5.6 model family and Work tool

OpenAI Unveils GPT-5.6 Model Suite and Work Tool

OpenAI releases GPT 5 6 model - OpenAI, a leading AI firm, has unveiled the GPT-5.6 model family, signaling the conclusion of a phased release strategy influenced by U.S. government guidelines. This move comes amid concerns about safety mechanisms in recent advanced models, which prompted cautious distribution prior to the full launch.

The new models are divided into three distinct categories—Sol, Terra, and Luna—each optimized for different functions. Sol is positioned as the most robust variant, while Luna focuses on efficiency. Terra, on the other hand, aims to balance performance and resource usage for routine tasks. OpenAI positions these models as versatile tools, emphasizing reliability alongside improvements in agentic coding and judgment.

Complementing the model release, OpenAI introduced the ChatGPT Work agentic AI tool. The company claims this tool can "maintain focus on complex projects for hours," offering sustained support for tasks requiring prolonged attention. It operates by distributing responsibilities among subagents, enhancing productivity for intricate workflows.

ChatGPT Work integrates with multiple platforms, including web services, mobile devices, and desktop systems. It can compile information from interconnected apps and files to generate documents, spreadsheets, and presentations. The tool’s adaptability extends to real-time collaboration, enabling seamless updates across messaging platforms like Microsoft Teams and Slack.

Examples include the tool’s ability to independently convert chat logs from Teams and Slack into structured documents or slides, then notify teams of critical changes. This functionality is now available in the Mac and Windows versions of ChatGPT apps, with a web interface expected to launch soon.

Anthropic’s Model Suspension Amid U.S. Export Controls

In a parallel development, Anthropic temporarily halted access to its Fable 5 and Mythos 5 models to align with a U.S. government directive. The decision followed alerts from the Commerce Department about a "jailbreaking" method that could circumvent safety protocols in Fable 5, potentially exposing software vulnerabilities.

"The directive restricts foreign nationals, including non-U.S. employees, from interacting with the models," stated a spokesperson. "This pause ensures compliance with emergency export regulations."

Anthropic’s action reflects broader regulatory efforts to manage AI risks, as the U.S. seeks to control the spread of advanced technologies. The suspension underscores the importance of safeguarding against unauthorized use of frontier models.

Vishal Mathur serves as Technology Editor for Hindustan Times. When not analyzing technological trends, he explores the intersection of analog and digital experiences in modern innovation.