OpenAI on Thursday unveiled GPT-5.6, a new family of models that raises the stakes in an increasingly crowded artificial intelligence market.
Three Models, Three Price Points
GPT-5.6 launches with three variants: Sol, the flagship workhorse; Terra, a mid-tier option; and Luna, a lower-cost model designed to broaden access. OpenAI says the lineup is built to expand what users can do across enterprise workflows, software development and scientific research.
Follow THE FUTURE on LinkedIn, Facebook, Instagram, X and Telegram
Chief Executive Sam Altman has argued that the new generation is dramatically more efficient and cost-effective than earlier versions. In a recent CNBC interview, he said Sol is 54% more token efficient for AI coding tasks, underscoring the company’s push to make performance and economics its key differentiators.
A Stronger Push Into Cybersecurity
OpenAI is also making a notable claim about security: it describes GPT-5.6 as its “strongest cybersecurity model yet,” saying it achieves frontier performance while using significantly fewer tokens.
That positioning comes against a sensitive backdrop. The model’s cyber capabilities drew attention after the Trump administration reportedly sought to restrict its rollout over concerns about misuse. OpenAI says GPT-5.6 is designed to support defensive work, including threat modeling, code review and patching, as well as blue teaming, the practice of simulating attacks to expose vulnerabilities before adversaries can exploit them.
ChatGPT Work Targets The Office Stack
Alongside the model launch, OpenAI introduced ChatGPT Work, a workplace-focused tool for enterprise teams. Available on desktop, web and mobile, it is designed to assist with routine clerical tasks such as drafting documents, building spreadsheets and preparing presentations.
The move reflects a broader industry trend: AI vendors are no longer competing only on raw model capability, but on whether their products can become part of the daily operating system for knowledge workers.
Anthropic Remains The Benchmark To Beat
OpenAI’s latest release arrives amid a wave of new model announcements from rivals, including SpaceXAI and Meta. But the company’s messaging appears aimed most directly at Anthropic, which has earned a strong reputation among enterprise buyers and emerged as one of OpenAI’s most credible competitors.
To make its case, OpenAI points to the Artificial Analysis Coding Agent Index, saying GPT-5.6 outperforms Anthropic’s models across the board. The company says Sol is its “best coding model yet,” and compares it directly with Anthropic’s recently released Fable.
According to OpenAI, Sol sets a new state of the art at 80, beating Fable 5 by 2.8 points while using less than half the output tokens, taking less than half the time and costing roughly one-third less. The company says Terra performs just above Fable 5, while Luna outperforms Opus 4.8.
Pricing And Availability
GPT-5.6 is now available across ChatGPT, Codex and the OpenAI API. Pricing per million tokens is set at $5 input and $30 output for Sol, $2.50 input and $15 output for Terra, and $1 input and $6 output for Luna.
For OpenAI, the launch is not just another product update. It is a calculated attempt to reinforce its lead in frontier AI while pressing its advantage in the enterprise, coding and cybersecurity markets where the next phase of competition is likely to be won.
