Anthropic has released two versions of its latest Claude model: one for approved users and another for everybody else.
Claude Fable 5.1 is generally available, while Claude Mythos 5.1 is offered “only through trusted access programs,” according to an Anthropic blog post. They share a model but use different safeguards.
The two-tiered structure comes as Silicon Valley developers have split over how best to shore up safety as artificial intelligence grows ever more powerful.
According to Anthropic, Mythos is designed for cybersecurity and life-sciences work, with access limited to vetted cyberdefenders and life scientists at U.S. organizations, though the company says it is “coordinating with the US government to expand access to a broader set of domestic and international partners as quickly as possible.”
The split sharpens a central DeAI question: Who decides which users may wield the most capable intelligence? Anthropic is not merely keeping the AI’s inner workings secret. It is putting the model’s least constrained capabilities behind an institutional gate.
DeAI News reported in July that more than 75 technology companies had backed Nvidia CEO Jensen Huang’s call to promote open-source models, including OpenAI and Google, while Anthropic remained the notable holdout. CEO Dario Amodei has argued that powerful models should undergo mandatory safety testing because of cyber and biological risks.
The issue came to the fore in June, when Anthropic took both Mythos 5 and Fable 5 offline after a Trump administration directive barred foreign nationals from using them.
Fable 5.1 loosens some restrictions, including those on using the model to discover software vulnerabilities, but not to exploit them. Anthropic says Claude Code users should see about 60% fewer safeguard interventions per session than with Fable 5.
Anthropic provided a link where users can “register interest in access” to Mythos 5.1 through its Cyber Verification Program, which the company says will add Mythos-class models “in the near future.”
HOW AI WAS USED IN THE PRODUCTION OF THIS PIECE: I created the initial version of this story using a custom skill designed to write story drafts based on blog posts. I filed the piece directly to DeAI News on Distro using the Distro Publisher MCP connector, and then I edited the piece (and rewrote substantial parts of it) using our built-in Distro editing interface. I checked the facts prior to publication. I created the story illustration on ChatGPT using a custom project.