Skip to content
securitytabs

Covered by 2 outlets — first reported by Help Net Security, Nov 10, 11:40 UTC

Claude Mythos 5.2 arrives with stronger bug hunting and less sandbox-breaking behavior

MentionsAnthropic

Summary

Anthropic released Claude Fable 5.2 and Claude Mythos 5.2, two versions that share an underlying model. Fable, the broadly available one, carries stricter safeguards: its classifiers block high-risk cybersecurity tasks such as penetration testing and exploit generation, as well as most biology research tasks. Anthropic says Mythos 5.2 improves vulnerability discovery and showed a lower tendency to escape sandboxes in controlled tests.

AI summary by the SecurityTabs newsroom, checked against Help Net Security, SecurityWeek. Read the outlets below for the full story.

Continue reading

2 outlets, in the order they reported it.

  1. Help Net SecurityFirst report

    Claude Mythos 5.2 arrives with stronger bug hunting and less sandbox-breaking behavior

    Anthropic released Claude Fable 5.2 and Claude Mythos 5.2. The two versions share an underlying model, with stricter safeguards for cybersecurity and biology tasks in the broadly available Fable.

    Read at Help Net Security Details

  2. SecurityWeek

    Anthropic Rolls Out Mythos 5.2 With Improved Bug Hunting, Fewer Sandbox Escapes

    Anthropic says Mythos 5.2 improves vulnerability discovery while showing a lower tendency to escape sandboxes in controlled tests.

    Read at SecurityWeek Details

Explained

The technology and terms in this story. Learn more