US tightens AI controls; Anthropic partially lifts restrictions, OpenAI opts for phased release

The United States is intensifying its oversight of cutting-edge artificial intelligence, with recent actions affecting leading developers Anthropic and OpenAI. The Commerce Department has partially lifted export controls on Anthropic's advanced AI model, "Mythos 5," allowing its provision to certain trusted partners. This move comes after initial restrictions were imposed in April due to national security concerns, leading Anthropic to suspend services globally for "Mythos 5" and "Fable 5."

While "Mythos 5" is now accessible to members of Anthropic's "Project Glasswing," the core "Fable 5" model, which had safety features applied before its general release, remains under tighter control. Details regarding any additional safety measures implemented by Anthropic in response to the eased restrictions have not been disclosed.

“

It's not a bad idea in itself, but I don't like the government deciding who gets to use cutting-edge models.

Meanwhile, OpenAI has opted for a phased rollout of its latest model, "GPT-5.6," in response to government requests. The model, available in three versions – "Sol," "Terra," and "Luna" – sees its top-tier "Sol" version reportedly outperforming Anthropic's "Mythos 5" in coding benchmarks. However, OpenAI's CEO, Sam Altman, voiced his discomfort with the government dictating which clients can access advanced AI models, despite acknowledging the value of safety testing.

These government interventions stem from an executive order signed earlier this month, mandating AI developers to voluntarily submit high-performance models for government review up to 30 days before public release. Experts anticipate escalating debates over balancing AI safety with industrial competitiveness, with concerns that prolonged institutional frameworks for AI releases could allow China to narrow the technological gap.

“

The government's move is a realistic transitional response. The longer the period before institutional frameworks are established for U.S. companies to widely release new models to the market, the greater the possibility for China to narrow the technology gap.