OpenAI Unveils GPT-5.6 AI Model Family With New Enterprise and Coding Tools

OpenAI has introduced GPT-5.6, a new family of artificial intelligence models designed to support enterprise applications, software development and scientific research. The release includes three models, Sol, Terra and Luna, each targeting different performance requirements while expanding the company's AI offerings across ChatGPT, Codex and the OpenAI API.

The launch comes as competition among AI developers continues to intensify, with technology companies racing to deliver faster, more capable and cost-efficient models for businesses and developers. OpenAI says the GPT-5.6 family focuses on stronger coding performance, improved efficiency and enhanced cybersecurity capabilities.

Three Models Designed for Different Workloads

According to OpenAI, Sol serves as the flagship model for demanding workloads, Terra is intended for mid-tier applications, and Luna provides a lower-cost option for customers seeking affordable AI deployment. CEO Sam Altman recently said that  Sol is 54% more token-efficient than previous models for AI coding tasks.

The company also describes GPT-5.6 as its strongest cybersecurity model to date. It supports defensive security activities including threat modelling, code review, software patching and blue team exercises, allowing organisations to identify potential vulnerabilities before they can be exploited.

Alongside the new models, OpenAI introduced ChatGPT Work, a workplace assistant available on desktop, web and mobile. The tool is designed to help enterprise users create documents, spreadsheets and presentations while assisting with routine business tasks.

OpenAI Highlights Coding Performance

OpenAI is positioning GPT-5.6 as a major upgrade for developers. Using results from the Artificial Analysis Coding Agent Index, the company says Sol outperformed Anthropic’s Fable 5 while using fewer than half the output tokens, completing tasks in less than half the time, and operating at roughly one-third the cost.

The company also said Terra scored above Fable 5 on the same benchmark, while Luna outperformed Opus 4.8. These comparisons form part of OpenAI’s effort to demonstrate the competitiveness of its latest models in enterprise software development.

Available Across OpenAI Platforms

OpenAI confirmed that the GPT-5.6 family is immediately available through ChatGPT, Codex and the OpenAI API. Pricing begins at $1 per million input tokens and $6 per million output tokens for Luna, while Terra is priced at $2.50 per million input tokens and $15 per million output tokens. Sol, the highest-performance model, costs $5 per million input tokens and $30 per million output tokens.

The release follows several recent AI announcements from competing technology companies, reflecting the industry’s continued focus on enterprise customers seeking advanced coding tools, productivity assistants and secure AI systems. With GPT-5.6 and ChatGPT Work, OpenAI is expanding its portfolio while emphasising performance and enterprise-ready capabilities across its latest generation of AI models.