OpenAI Unveils New GPT-5.6 AI Model Lineup with Enhanced Capabilities

OpenAI has unveiled its latest lineup of AI models, GPT-5.6, featuring three variants: Sol, Terra, and Luna. Designed for diverse use cases—from corporate tasks to scientific research—Sol serves as the flagship version, Terra as the mid-tier, and Luna as the budget-friendly option. According to CEO Sam Altman, the new models significantly surpass previous versions in efficiency and cost-effectiveness, particularly in programming-related tasks.

The announcement highlights GPT-5.6’s advanced cybersecurity capabilities. OpenAI describes the new lineup as the “most powerful data protection model,” capable of tasks like threat modeling, code audits, and attack simulations (blue teaming). These features enable companies to identify vulnerabilities in their systems before malicious actors can exploit them. The model’s launch has previously raised concerns among regulators due to potential misuse risks.

OpenAI also introduced a new tool, ChatGPT Work, aimed at enterprise users. Available on desktop, web, and mobile devices, it helps automate routine tasks such as document creation, spreadsheets, and presentations. The GPT-5.6 models are now accessible via ChatGPT, Codex, and the OpenAI API. Pricing varies: Sol costs $5 per million input tokens and $30 per million output tokens, Terra is priced at $2.50/$15, and Luna at $1/$6.

The new releases come amid competition from rivals like SpaceXAI and Meta*. However, Anthropic remains OpenAI’s primary competitor, having gained traction among corporate clients. OpenAI claims its models outperform Anthropic’s offerings in key benchmarks, including the Artificial Analysis Coding Agent index. For instance, Sol scores 80 points, surpassing Anthropic’s Fable 5 by 2.8 points while using fewer resources and time.