OpenAI GPT-5.6 Sol: Trump-Approved Release Amid Cyber Review

OpenAI is limiting the release of its new artificial intelligence model, GPT-5.6 Sol. It will only be available to a small group of trusted partners approved by the Trump administration. This restriction comes at the request of the government, which is conducting a cybersecurity review of AI products. OpenAI stated it doesn't believe this government access process should become the long-term default, viewing it as a temporary step before broader availability. This follows similar actions against OpenAI rival Anthropic, which had to take two AI models offline due to a Trump directive. The White House is concerned about powerful AI models, especially after Anthropic warned its Mythos model could find software flaws that malicious hackers could weaponize. OpenAI says its Sol model is better at helping fix vulnerabilities than carrying out cyberattacks, but acknowledges potential unforeseen risks, leading to this phased release. This shows how government oversight is directly impacting the public availability of new AI technologies.