OpenAI and Anthropic, the two leading artificial intelligence laboratories, have jointly urged Washington to create a national framework for evaluating frontier AI models. The Trump administration ordered in June 2026 a voluntary 30-day review process for advanced models before public release. In response, OpenAI restricted access to its GPT-5.6 Sol model and Anthropic did the same with several Claude variants, including Mythos 5. Both companies specifically warned about national security risks linked to Chinese open-weight AI models. Anthropic further called publicly for a slowdown in artificial intelligence advancements, particularly regarding recursive self-improvement.
Source: Read the original article
