OpenAI GPT-5.6 model series is moving closer to wider availability, as the company prepares to expand access to its latest generation of artificial intelligence systems after an initial limited preview. The new lineup includes GPT-5.6 Sol, GPT-5.6 Terra and GPT-5.6 Luna, three models designed for different levels of performance, speed and cost.
OpenAI first introduced the GPT-5.6 family in late June, describing Sol as its flagship and most capable model, Terra as a balanced model for everyday work, and Luna as a faster, lower-cost option. According to the company, the new series advances performance across software engineering, computer use, professional knowledge work, scientific research and cybersecurity. The launch also adds to wider debates around large language models, especially as more advanced systems become part of daily work, research and decision-making.
The rollout has attracted attention because of the models’ stronger ability to work with software code and identify security weaknesses. These capabilities could help developers, researchers and cyber defenders find vulnerabilities and create fixes more efficiently. At the same time, they have raised concerns about possible misuse if similar tools are used to search for exploitable flaws in software systems.
GPT-5.6 Sol, Terra and Luna Explained
The GPT-5.6 family is built around three separate model tiers. Sol is the most powerful option in the series and is positioned as OpenAI’s new flagship model. Terra is designed as a mid-range model for more common professional and productivity tasks, while Luna is the fastest and most affordable option for users who need lower-cost performance.
OpenAI says the naming system is meant to give developers and organizations clearer choices across intelligence, speed and price. During the preview period, access has been limited to selected trusted partners and organizations through the OpenAI API and Codex, rather than being available to all ChatGPT users.
Pricing has also been outlined for the new model family. GPT-5.6 Sol is priced at $5 per 1 million input tokens and $30 per 1 million output tokens. Terra is priced at $2.50 per 1 million input tokens and $15 per 1 million output tokens, while Luna costs $1 per 1 million input tokens and $6 per 1 million output tokens.
Why the GPT-5.6 Launch Is Drawing Cybersecurity Attention
One of the biggest discussion points around the OpenAI GPT-5.6 model series is cybersecurity. OpenAI says GPT-5.6 Sol is its most capable model so far for security-related work, including vulnerability research and software analysis. The company also says the model is better at helping users find and fix weaknesses than reliably carrying out full end-to-end attacks.
Because of these capabilities, OpenAI has added stronger safeguards to the new model family. These include model-level protections, real-time checks, account-level monitoring, differentiated access and continued testing during the preview period. The company says these safeguards are intended to make prohibited offensive use more difficult while still allowing legitimate security work such as code review, debugging, patch development, security education and defensive testing.
OpenAI has also said GPT-5.6 Sol does not cross its Cyber Critical threshold under the company’s preparedness framework. However, the company acknowledged that no evaluation can capture every possible real-world use of a powerful model, especially when it can be combined with other tools or used across longer workflows.
Government Review and Limited Preview
OpenAI said it shared access to GPT-5.6 with a small group of trusted partners as part of a phased release process. The company also said it previewed its plans and the models’ capabilities with the US government ahead of launch.
According to OpenAI, this approach is intended to allow more time for testing and coordination before broader availability. The company has said it does not believe this kind of government access process should become the long-term default, but described the limited preview as a short-term step toward a broader rollout.
Reports have also suggested that US officials approved a wider release of GPT-5.6 after testing and discussions with OpenAI. However, OpenAI’s own help documentation states that the company has not announced a general-availability date and plans to expand access as soon as possible.
What This Means for ChatGPT and Developers
For now, GPT-5.6 is not generally available to individual ChatGPT users during the preview period. Approved organizations may access Sol, Terra and Luna through the OpenAI API, Codex or both, depending on their approval.
The model family is expected to matter most at first for developers, enterprise teams and technical users working on code, research, automation and cybersecurity tasks. Over time, similar advances could also influence AI tools for digital marketing, customer support, content production and other business workflows. If OpenAI expands access more broadly, GPT-5.6 could become one of the company’s most important public AI releases, especially because of its stronger reasoning, coding and security capabilities.
The timing of the wider launch will be closely watched by the technology industry. With competing AI companies also preparing more powerful models, OpenAI’s GPT-5.6 series could play a major role in the next phase of competition around advanced artificial intelligence systems.