U.S. to Introduce Voluntary AI Model Release Standards, Affecting OpenAI and Anthropic

ChainCatcher report, according to the Financial Times, the U.S. government is in discussions with multiple AI companies to introduce, as early as next week, a voluntary industry standard for the release of advanced AI models to prevent the misuse of cutting-edge technologies by foreign nations. The standard will establish clear guidelines on model performance benchmarks, release timelines, and access permissions both domestically and internationally. Due to recent tightening of regulatory policies, several leading AI companies have already adjusted their operations. OpenAI has delayed the full release of GPT-5.6 at the government’s request, making it available only to a limited number of vetted partners; Anthropic’s two top models, which faced export restrictions for nearly three weeks, had their controls lifted this week; Google is also in close communication with the government as it prepares its next-generation code model. Additionally, the report notes that both OpenAI and Anthropic, currently under regulatory scrutiny, are actively preparing for their initial public offerings (IPOs).