Early users of Anthropic’s Mythos AI model are maintaining their access to a preview version, even amid a US government shutdown order that has restricted access to other models. This continued access is part of Project Glasswing, which includes around 200 organizations, such as banks and technology firms, that were granted permission by Anthropic PBC to utilize Mythos for identifying cyber vulnerabilities. Notably, firms like Cisco Systems Inc. and Dragos Inc. are among those with preserved access, despite the government’s recent export control directive aimed at limiting advanced model availability to foreign nationals due to national security concerns.

Anthropic: Anthropic is an AI safety and research company focused on building reliable, interpretable, and steerable artificial intelligence systems, including its Claude family of models. It recently released advanced models such as Fable 5 and Mythos 5 before receiving a US government export control directive. The company responded by suspending access to these models for all users to comply with restrictions aimed at foreign nationals.
Mythos AI: Mythos AI is Anthropic’s advanced frontier model, referred to as Mythos 5, designed for high-capability tasks in areas like coding, vision, and long-context processing. It was directly impacted by the US government’s recent national security order limiting access. Despite the broader suspension, early users have retained access to a preview version of the model.

`json
{
“Access”: “Early users continue to access Anthropic’s Mythos AI model’s preview version despite the US government shutdown order.”,
“Regulation”: “The US government issued an export control directive to Anthropic requiring suspension of access to its most advanced models over national security concerns.”
}
`