US Government Restricts Access to OpenAI's Latest AI Model (heise.de)

• OpenAI restricts access to GPT-5.6 preview at US government request.
• The 'Sol' variant can autonomously handle programming, biology, and cybersecurity tasks.
• Initial access limited to US partners; international rollout planned.
• Similar forced withdrawal of Anthropic's Fable 5 model over cyber misuse fears.

"The US government has compelled OpenAI to restrict access to the preview version of its GPT-5.6 AI model to a small group of trusted partners, citing concerns about potential misuse for cyberattacks. OpenAI stated that the restricted release was requested by the government, with initial access limited to US entities only. The company is working to extend access to international partners in the coming weeks, while emphasizing that such controls should not become the norm."