[Forbes: Partial Lifting of Export Restrictions on Anthropic Models in the U.S. Signals 'Tiered Relaxation' Phase for AI Regulation] Golden Finance reports that Forbes has published an article stating that the U.S. Department of Commerce has partially lifted export restrictions on Anthropic's advanced AI models, allowing specific controlled users to regain access to its cybersecurity model Claude Mythos 5. However, another advanced model, Fable 5, remains restricted, with related discussions still ongoing. Industry analysts point out that this policy adjustment, which combines restrictions with relaxation, signals that AI regulation is entering a 'tiered relaxation' phase. It also reflects the complex balance between national security, data sovereignty, and international competition concerning advanced AI models. Against the backdrop of accelerating global AI competition, models with vulnerability detection capabilities are facing stricter scrutiny, and countries are beginning to discuss cross-border regulatory frameworks for advanced AI capabilities.