Leaders at a G7 meeting discuss ways to allow “trusted partners” access to cutting-edge US artificial intelligence, according to multiple reports. The debate follows changes by Anthropic, which is described as disabling access to its models Fable 5 and Mythos 5. The articles say the G7 is considering a route that would enable use by non-US parties despite restrictions on non-American deployment or access to these advanced AI systems. The reports do not specify the exact mechanism being considered, but characterize the discussion as seeking a compromise between regulatory or licensing limits applied to US-developed AI and the needs of partner countries. All outlets frame the issue around access and eligibility for advanced AI models, with Anthropic’s reported access changes acting as a trigger for the G7 discussion. No further technical details, participant lists, or policy commitments are provided in the excerpts.