.jpg)
Just days after expanding its cybersecurity-focused frontier AI program, Project Glasswing, Anthropic released two new models, Claude Mythos 5 and Claude Fable 5.
Mythos 5 has been presented by Anthropic as “an upgrade to Claude Mythos Preview” and the AI company claimed in a June 9 announcement that it has “the strongest cybersecurity capabilities of any model in the world.”
This new upgrade will initially be deployed through Project Glasswing, in collaboration with the US government, but the company plans for a wider availability of the model in the future via “a broader trusted access program.”
Meanwhile, Fable 5 has is said to be powered by the same underlying frontier AI model as Mythos 5 but with additional guardrails, especially in areas like cybersecurity where the company said it “could be misused to cause serious damage.”
One of these safeguards involves a system where queries made by Fable 5 on certain topics will receive a response from Claude Opus 4.8, Anthropic’s next-in-line model that is available to everyone.
“To release the model both safely and quickly, we’ve tuned these safeguards conservatively – they’ll sometimes catch harmless requests, though they trigger, on average, in less than 5% of sessions,” cautioned the AI company, adding that is it working on new models where these safeguards will be fine-tuned to reduce false positives.
Fable 5 and Mythos 5 are being offered at $10 per million input tokens and $50 per million output tokens, less than half the price of Claude Mythos Preview at $25 per million input tokens and $125 per million output tokens.
Quickly following the Anthropic announcement, Microsoft said Fable 5 is now available in Microsoft Foundry and can be used to power agents in GitHub Copilot and Foundry Agent Service.
Fable’s Guardrails May Not Be Sufficient, Cyber Experts Warn
The Anthropic announcement generated mixed feelings in the cybersecurity industry.
Sachin Puri, CEO of Network Solutions, said Mythos-class models “represent a meaningful step forward in how AI can help identify, analyze and respond to increasingly sophisticated cyber threats and detect vulnerabilities.”
However, he warned that, as these systems become more powerful, their use must “include governance, accountability and what happens when guardrails fail.”
Jamie Moles, senior technical manager at ExtraHop, expressed skepticism for the need to launch a “Mythos-class” tool to the public after widespread warnings from safety experts that the model poses severe risks to corporate security.
He said he hoped Anthropic and any company developing powerful AI models for mass deployment deploys these innovations responsibly “to ensure its technology does not compromise the stability of systems the public relies on every day.”
Andrew Rubin, founder and CEO of Illumio, is even more brazen, warning that the introduction of guardrails isn’t evidence that the problem is solved.
“It’s an admission that even the companies building these models don’t fully trust where the capability leads,” he said.
“Constraints at the interface don’t change the underlying math. Attackers won’t operate at that layer. They’ll go straight after the capability itself. And as these tools become more broadly available, the speed and scale of attacks will only increase. The real question is whether defenders are prepared to operate at the same speed.”
Douglas McKee, director of vulnerability intelligence at Rapid7, believes that Fable will not replace Mythos and will instead become “the everyday workhorse” while Mythos remains “the power tool for tightly vetted security and infrastructure teams.”
Therefore, he argued that Anthropic must “keep getting Mythos-class access into the hands of vetted defensive teams through restricted channels and its partner ecosystem” to ensure cyber defenders “keep pace with increasingly capable threat actors.”
Image credits: Stockinq / PhotoGranary02 / Shutterstock.com
.jpeg)
