The News
Anthropic on Tuesday launched a version of its powerful unreleased Mythos model that the company said was safe for general use. Fable 5 has guardrails in place that Anthropic said prevent it from answering questions about cybersecurity and biology — capabilities that the company said made Mythos too dangerous for public release.
The big question is whether the guardrails for Fable 5 are strong enough to withstand jailbreaking attempts by bad actors who Anthropic expects will flock to the model after its high-profile warnings about Mythos’ abilities to upend cybersecurity. The company said it “extensively” tested the model with hackers who tried to bypass its safeguards, and none were successful. Instead, Anthropic’s less powerful Opus 4.8 model stepped in to answer those questions.
Anthropic acknowledged that without safeguards, Fable 5’s capabilities could be misused in dangerous ways. “The difference [in Fable and Mythos] is not what the model can do, but what our safeguards will allow,” a spokesperson told Semafor. “Fable 5, without safeguards, would be exceptionally strong at finding and exploiting software vulnerabilities, which could meaningfully lower the cost of cyberattacks.”
In early testing, customers claimed Fable 5 significantly reduced the time it took to publish software and performed well on reasoning tasks.
Know More
On Tuesday, Anthropic also rolled out an upgraded version of Mythos to select customers who were previously granted access. Mythos 5 “has the strongest cybersecurity capabilities of any model in the world,” the company said.
Both Mythos 5 and Fable 5 are priced lower than the previous version of Mythos, though the long-running analytical tasks make them more expensive than Anthropic’s other models.