OpenAI Unveils New GPT-5.6 Model Family, Emphasizing Cybersecurity and Enterprise Prowess

OpenAI unveiled its newest family of AI models, GPT-5.6, marking a significant entry into an increasingly crowded and competitive field. This family comprises three distinct variants: Sol, positioned as the robust workhorse; Terra, an intermediate option; and Luna, designed for cost-effectiveness. These models are engineered to significantly broaden user capabilities across diverse sectors, with OpenAI highlighting their potential in demanding enterprise operations, advanced coding tasks, and complex scientific research. This launch follows similar announcements from rivals like SpaceXAI and Meta this week, intensifying the race for AI dominance. CEO Sam Altman has underscored the new models' superior efficiency and cost-effectiveness compared to their predecessors, claiming they are orders of magnitude better. He noted that Sol, for instance, exhibits 54% greater token efficiency specifically for AI coding tasks. A standout feature of the GPT-5.6 family is its advanced cybersecurity capabilities. OpenAI touts 5.6 as its "strongest cybersecurity model yet," achieving frontier performance with remarkably fewer tokens. This aspect has garnered significant attention, especially given past concerns from the Trump administration regarding potential misuse, which had sought to restrict its initial rollout. The GPT-5.6 models are specifically designed to bolster defensive cybersecurity activities. Their applications include sophisticated threat modeling, comprehensive code review and patching, and blue teaming exercises—simulating attacks on internal systems to identify vulnerabilities before malicious actors can exploit them. Complementing these powerful models, OpenAI also introduced ChatGPT Work, a new enterprise-focused tool. This workplace companion, accessible via desktop, web, and mobile, is tailored to assist teams with daily clerical tasks, such as drafting documents, managing spreadsheets, and creating presentations, streamlining office workflows. OpenAI's latest launch appears strategically aimed at its primary competitor, Anthropic, which has successfully positioned itself as an "underdog" focusing on enterprise clients and gaining a growing share of support. OpenAI directly challenges Anthropic's market position by citing the Artificial Analysis Coding Agent Index, a recognized benchmarking metric. The company asserts that its new family of models consistently outperforms Anthropic's offerings across the board. Sol, specifically hailed as OpenAI's "best coding model yet," is explicitly compared to Anthropic's recently released and much-hyped Fable. According to OpenAI's data from the Coding Agent Index, Sol achieves a new state-of-the-art score of 80, surpassing Fable 5 by 2.8 points while using less than half the output tokens, requiring less than half the time, and costing approximately one-third less. This performance advantage extends throughout the family, with Terra reportedly performing just above Fable 5 and Luna outperforming Opus 4.8. The GPT-5.6 family is now widely available across ChatGPT, Codex, and the OpenAI API. Pricing per million tokens is set at $5 input / $30 output for Sol, $2.50 input / $15 output for Terra, and $1 input / $6 output for Luna, offering various options to suit different user needs and budgets.