US Government Eases Restrictions on Anthropic's Mythos 5 AI Model Following Tense Negotiations

The Trump administration has partially eased restrictions on Anthropic, allowing the artificial intelligence safety startup to restore access to one of its most powerful models for a select group of domestic organizations. The decision, delivered in a letter from the Commerce Department, marks a significant de-escalation in a high-stakes standoff between the White House and the San Francisco-based AI developer.

The Export Control Clash and Sudden Shutdown

The conflict erupted in mid-June 2026 when the Trump administration issued an abrupt export control directive. This order forced Anthropic to pull its cutting-edge models, Mythos 5 and Fable 5, offline and disable access for foreign nationals. The sudden block was triggered by cybersecurity and national security concerns, treating the frontier AI systems similarly to restricted military technology. Shortly before the directive was issued, President Donald Trump commented on the strategic importance of the sector, stating that the government was holding meetings with the industry and asserting that whoever leads in AI will lead the world.

Negotiating in a Regulatory Vacuum

Following the shutdown, Anthropic and the White House engaged in days of tense negotiations. The dispute highlighted a highly unpredictable regulatory landscape. While the Trump administration had actively dismantled Biden-era national AI frameworks and established a federal task force to challenge state-level AI regulations, it simultaneously clamped down on Anthropic. A person close to the company revealed that Anthropic believed it had not violated any concrete, established rules. Conversely, the White House contended that the AI lab had acted recklessly, proving it could not be trusted to deploy frontier systems safely without oversight.

A Partial Resolution for Mythos 5

The breakthrough came in a letter sent by Commerce Secretary Howard Lutnick. In the correspondence, Lutnick noted that Anthropic had worked closely with the federal government to address risks associated with the restricted models, stating that these collaborative efforts had yielded significant progress. Under the revised terms, Anthropic is permitted to restore access to the Mythos 5 model for an approved list of U.S. companies and organizations. Additionally, Anthropic has committed to establishing cooperative protocols with the government for all future model releases.

Despite this progress, the restrictions have not been entirely lifted. Negotiations are still ongoing to restore access to Fable 5 and to expand the distribution of Mythos 5 to a broader range of organizations. An Anthropic spokesperson expressed satisfaction with the recent developments, stating that the company is pleased with the progress and remains committed to working alongside federal authorities. This episode underscores the reality that frontier AI companies must now navigate highly unpredictable, real-time government interventions even in an otherwise anti-regulatory political environment.