The Trump administration on Friday granted Anthropic permission to release its Mythos 5 model to roughly 100 companies and federal agencies, ending a two-week standoff that saw the AI startup disable access to both of its latest flagship models. The Commerce Department confirmed the decision in a letter addressed to Anthropic co-founder Tom Brown, who has reportedly taken over negotiations with the administration from CEO Dario Amodei. Secretary Howard Lutnick wrote: "I have determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model."

The Two-Week StandOff Explained

The drama began earlier this month when Anthropic announced its Fable 5 and Mythos 5 models to considerable fanfare, marketing them as state-of-the-art across multiple industry benchmarks. Within days, the company received an export control directive invoking "national security authorities" that required suspending all access by foreign nationals—including foreign national employees of Anthropic itself. Rather than comply with what it apparently viewed as an unworkable requirement, Anthropic disabled both models entirely rather than risk running afoul of the government order.

Fable 5 Remains in Limbo

The Mythos 5 approval doesn't extend to Fable 5, which remains inaccessible. This is significant because Fable 5 marked Anthropic's first public release of such an advanced model—a departure from its traditionally cautious approach to deployment. The company had developed new safeguards specifically designed to block responses in high-risk areas, positioning it as a responsible actor even while pushing the frontier of capability. That careful positioning apparently wasn't enough to satisfy regulators who remain concerned about foreign access to cutting-edge AI systems.

OpenAI Took a Different Approach

Meanwhile, Anthropic's chief rival OpenAI announced three new models—GPT-5.6 Sol, Terra, and Luna—on the same day, and disclosed it was complying with similar government requests by initially limiting rollout to a "small group of trusted partners." The key difference: OpenAI framed this as part of its normal launch strategy while working toward broader availability in coming weeks. Anthropic's situation has been messier, complicated by an ongoing dispute with the Department of Defense that predates this month's standoff.

The Pentagon Declared Anthropic a Supply Chain Risk

Earlier this year, the Department of Defense clashed with Anthropic over disagreements regarding use of its models in military applications. When negotiations collapsed, the DOD took the extraordinary step of declaring Anthropic a "supply chain risk"—a designation historically reserved for foreign adversaries. The label forces defense contractors to certify they won't use Claude models in their work with the military. Anthropic has sued the Trump administration attempting to reverse this blacklisting, and litigation remains ongoing.

Key Takeaways

  • Mythos 5 approved for ~100 partners; Fable 5 access still blocked by government order
  • Tom Brown now leading negotiations instead of CEO Dario Amodei
  • Anthropic disabled both models rather than comply with foreign national access restrictions
  • DOD previously blacklisted Anthropic as a supply chain risk—suit against administration ongoing
  • OpenAI navigated similar restrictions more smoothly, launching three new models same day

The Bottom Line

This partial victory for Anthropic masks deeper structural problems. The company that built its brand on safety and alignment now finds itself caught in Washington's crosshairs over export controls that may be fundamentally incompatible with how AI labs actually operate across borders. Mythos 5 getting a limited release doesn't resolve the underlying tension—and if Fable 5 stays blocked, Anthropic's competitive position takes a real hit.