During a working lunch with G7 leaders, international tech CEOs, and outreach partners focused on innovation and AI, Dario Amodei, the CEO of Anthropic, participated in discussions at the G7 Summit on June 17, 2026, held in Évian-les-Bains, France.
In a significant development, the U.S. government has officially permitted Anthropic to launch its Mythos 5 model to around 100 companies and federal agencies, as confirmed by CNBC.
The announcement came from the Commerce Department, which communicated a letter to Anthropic regarding this decision. This move represents a crucial advancement in the dialogue between the Trump administration and Anthropic, which has faced a two-week deadlock concerning its latest AI models, Fable 5 and Mythos 5.
U.S. Commerce Secretary Howard Lutnick stated in the letter, which was seen by CNBC, "I have determined that appropriate safeguards exist to allow certain trusted partners access to the Claude Mythos 5 Model."
While the letter was directed to Anthropic co-founder Tom Brown, it did not authorize the company to restore access to Fable 5, a task that Brown has taken on amid negotiations with the administration, stepping in for Amodei.
Earlier this month, Anthropic had to deactivate access to both Fable 5 and Mythos 5 in response to an export control directive from the government, which cited "national security authorities." The company clarified that this suspension applied to all foreign nationals, including its foreign employees, regardless of their location.
This move came just days following the public announcement of the new models, which were highlighted by Anthropic as leading-edge in various industry benchmarks. Fable 5 in particular marked a significant advancement as Anthropic's first high-profile public release, enhanced by new safety features aimed at preventing harmful responses in sensitive areas.
In a parallel development, rival firm OpenAI revealed three new AI models on the same day and confirmed its compliance with the government’s directive to initially limit access to a "small group of trusted partners."
In a blog post, OpenAI emphasized its commitment to providing widespread access, announcing plans to generally make models — GPT-5.6 Sol, Terra, and Luna — available in the upcoming weeks. They mentioned having shared their model capabilities and rollout plans with the government prior to the Friday announcement.
Anthropic has faced challenges with the Trump administration, having previously experienced tensions with the Department of Defense regarding the application of its models. Following stalled negotiations, the DOD labeled Anthropic as a supply chain risk, implying that the company poses a potential threat to U.S. national security, a designation typically reserved for international adversaries. This classification requires defense contractors to certify they will not utilize Anthropic’s Claude models for military applications.
In response, Anthropic has initiated legal action against the Trump administration in an effort to overturn its blacklisting, and the litigation remains active.
