Original article appeared here.
Anthropic has released its Claude Fable 5 artificial intelligence model for public use, alongside tighter safeguards for a more advanced variant known as Mythos 5.
The launch marks a shift in how one of the largest frontier AI providers handles access to its most capable systems. Rather than reducing the model's capabilities, Anthropic has separated product access and safety controls from the core model, offering different configurations for general users and higher-risk use cases.
Security experts say the move highlights the growing tension between rapid AI progress and the industry's ability to manage cyber risk. It comes as enterprise defenders report sustained attack volumes and regulators in multiple regions raise expectations for patching speed and incident response.
Several specialists argue that any reassurance from a “safe” version of Mythos is misplaced. They point to the speed at which advanced models can already discover and weaponise software vulnerabilities, and to the limited effectiveness of current guardrail and classification layers.
“The most honest thing Anthropic has done here is ship one model as two products. Splitting Fable 5 and Mythos 5 acknowledges that capability and safety are in genuine tension, and that pretending otherwise serves no one. But the most important line in the entire announcement isn't about the classifiers. It's buried in the operational detail: a high-severity vulnerability found by the model takes about two weeks to patch on average. Meanwhile, Mythos Preview built working exploits from a disclosed CVE in under a day. That gap is where risk lives. And no classifier closes it.
This makes concrete what the CSA data showed last week: enterprises aren't failing because they can't detect vulnerabilities. They're failing because they can't act on them fast enough. AI has collapsed the attacker's timeline to hours. The defender's timeline hasn't moved. Anthropic is right that the defensive head start only matters if the industry uses it. The harder truth is that most enterprises aren't yet equipped to - not because the tools don't exist, but because the governance architecture to deploy them safely hasn't kept pace with the capability. That's the real race,” said Gidi Cohen, Chief Executive Officer and Co-founder, Bonfy.AI.
Read the rest of the article here.