The U.S. government has permitted Anthropic to redeploy its Mythos 5 and Fable 5 AI models to a limited set of trusted U.S. partners [1, 2, 3, 5].
This decision follows a period of intense scrutiny over the potential for advanced artificial intelligence to pose systemic risks. By lifting these specific export curbs, the administration is balancing the need for national security with the desire to maintain a competitive edge in AI development within the United States [1].
The release comes after the Trump administration previously flagged the models as national-security risks [1]. Reports said the export curbs were lifted about three weeks after those initial security concerns were raised [1]. The process involved safety testing to ensure the models would not be misused or leaked to adversarial nations.
Earlier this month, reports said the government had begun granting the green light for a limited re-release of the Mythos 5 model [5]. The current authorization extends this permission to include the Fable 5 model as well [2]. These models are not being released to the general public, but are instead restricted to a curated group of companies and partners deemed trustworthy by federal regulators [3, 5].
Anthropic has worked with government officials to meet the necessary safety benchmarks to secure this approval [1]. The limited deployment allows the U.S. to monitor how these high-capacity models perform in real-world environments while preventing the technology from falling into the hands of foreign competitors [1, 5].
“The U.S. government has permitted Anthropic to redeploy its Mythos 5 and Fable 5 AI models.”
This move signals a shift toward a 'trusted partner' model of AI governance, where the U.S. government acts as a gatekeeper for the most powerful frontier models. By selectively lifting restrictions, the administration aims to foster domestic innovation and economic growth without compromising the security protocols designed to prevent the proliferation of dual-use AI capabilities to global rivals.
