At the World AI Conference (WAIC) in Shanghai, President Xi Jinping delivered a landmark keynote advocating for a people-centered 'symphony of international cooperation' in AI. The speech positioned China as a champion of open-source development and challenged the U.S. approach of 'overstretching' national security concepts. The event coincided with the launch of the 'Global AI Governance Rules Map 2.0,' an interactive tool to compare international regulatory frameworks, as the world's major powers compete to define the ethics and safety standards of frontier models.
A shift is occurring where the line between private labs and state infrastructure is blurring. With OpenAI reportedly negotiating a 5% stake for the US government and flagship models like GPT-5.6 and Claude Fable 5 facing White House-mandated delays for safety reviews, 'Sovereign AI' has evolved from a hardware goal into a governance requirement where frontier releases must now route through national security councils.
Moonshot AI's Kimi K3, released this week, is the world's largest open-weight model to date, featuring 2.8 trillion parameters.
Which flagship model tier did OpenAI officially release for public general availability this week after government-imposed delays?