OpenAI launched a new generation of model families, continuing to focus on business office, programming and network security scenes. GPT-5.6, which contains three versions of Sol, Terra and Luna, has been accessed through ChatGPT, Codex and OpenAI API.

Three versions overwhelm different prices

OpenAI locates Sol as the main force model, Terra as the middle version and Luna as the lower cost. According to the company, this series of models could be used for business workflows, code development and scientific missions.

The CEO of OpenAI, Sam Altman, had previously indicated that the new model represented a significant increase in efficiency and cost compared to previous generations. In a recent interview with CNBC, he stated that Sol had improved its efficiency by 54 per cent in the use of token in the AI coding task.

Network security capability is a priority

OpenAI describes GPT-5.6 as its strongest cybersecurity model to date and indicates that the model has reached frontier levels with less token consumption. According to the company, this capability is primarily oriented towards defensive uses, including threat modelling, code review and repair, and blue team exercises.

There has been much discussion about its network security capacity. The security applications of GPT-5.6 were also one of the priorities of the launch, according to reports that the Trump Government had attempted to limit the scope of the launch because of concerns about the misuse of the model.

Sync Up ChatGPTwork

In addition to model updates, OpenAI also released ChatGPTwork, which is located as an office assistant to the enterprise team and can be run on the desktop, web and mobile sides. This tool is used mainly for day-to-day matters such as drafting documents, preparing tables and presentations.

This release comes at a time when the competition for the AI Large Model continues to rise. Just this week, SpaceXAI and Meta launched similar products. OpenAI's advocacy focus is clearly on Anthropic, which has recently been expanding its influence through a corporate customer strategy.

OpenAI also compares Sol with Anthropic's recently released Fable. According to the company, on the basis of Coding Age Index, Sol scored 80, 2.8 points higher than Fable 5, while exporting token less than half of the latter, less than half the time, and about one third less.

  • Sol: Enter $5 per million token, output $30
  • Terra: Enter $2.5, output $15
  • Luna: Enter $1 and output $6