OpenAI unveiled on Thursday its new family of models, GPT-5.6, comprising three variants called Sol, Terra, and Luna, alongside the ChatGPT Work tool designed to assist enterprise teams with daily clerical tasks. The models are positioned as a response to growing enterprise demands, with a strong focus on cybersecurity and coding.
GPT-5.6 comes in three variants for diverse needs
The flagship model Sol is considered the workhorse, offering high performance at a cost of $5 per million input tokens and $30 per million output tokens. Terra is a mid-range option with halved prices, while Luna is the budget-friendly variant at $1 input and $6 output. CEO Sam Altman stated that Sol is 54% more token-efficient for coding tasks compared to previous versions. The models are now available on ChatGPT, Codex, and via API.
Sponsored Protocol
ChatGPT Work as an intelligent office assistant
Alongside the models, OpenAI released ChatGPT Work, an AI agent designed to automate clerical tasks such as drafting documents, spreadsheets, and presentations. It runs on desktop, web, and mobile, integrating with the browser for automatic scheduling. For more details, read the article on OpenAI Releases ChatGPT Work with Built-In Browser and Automated Task Scheduling. Additionally, the merger of Codex into ChatGPT has enhanced code generation capabilities, as described in OpenAI Merges Codex into ChatGPT and Launches Work Agent with GPT-5.6 Sol and Terra Models.
Sponsored Protocol
Direct comparison with Anthropic
OpenAI explicitly compared its new family with Anthropic's models, particularly Fable 5. According to the Artificial Analysis Coding Agent Index, Sol scores 80, surpassing Fable 5 by 2.8 points while using less than half the output tokens, less than half the time, and costing about one-third less. Terra also ranks just above Fable 5, and Luna outperforms Opus 4.8. This positioning aims to counter Anthropic's growing share in the enterprise segment.
Cybersecurity at the core of new features
One of the most notable advancements is the emphasis on cybersecurity. OpenAI calls GPT-5.6 its strongest cybersecurity model yet, achieving frontier performance with significantly fewer tokens. It supports defensive activities such as threat modeling, code review and patching, and blue teaming, which simulates attacks to identify vulnerabilities. The Trump administration had previously sought to restrict the model's rollout due to misuse concerns, but OpenAI proceeded, ensuring the capabilities are defense-oriented. For an in-depth analysis, refer to the original TechCrunch article: OpenAI launches its new family of models with GPT-5.6.
Sponsored Protocol
With GPT-5.6 and ChatGPT Work, OpenAI aims to solidify its position in the enterprise AI market, offering more efficient and secure tools for businesses of all sizes.
Source: https://techcrunch.com/2026/07/09/openai-launches-its-new-family-of-models-with-gpt-5-6
