For practitioners, a public release of a frontier model after direct government review raises the bar for red-teaming, vulnerability scanning, and deployment controls across organizations that use large language models. According to AFP via Yahoo Finance, OpenAI posted on X: "GPT-5.6 Sol, along with Terra and Luna, will launch publicly this Thursday. We're expanding preview access globally now." AFP reports OpenAI had shared preview access to GPT-5.6 with a limited group of US-based partners at Washington's request. Axios, cited by AFP, reported the US administration gave the company a green light for a broader launch following technical testing and meetings with officials. AFP also notes similar concern about Anthropic's Mythos series and the models' ability to identify software vulnerabilities.
OpenAI's GPT-5.6 launches Thursday after a delay forced by the U.S. government