Skip to content
Brief Security & Trust ·

Anthropic ships Claude Fable 5.2 and Mythos 5.2, one model with two cyber settings

Fable 5.2 blocks fewer harmless requests, while Mythos 5.2 finds more software flaws in Anthropic's tests.

In brief
  • Fable 5.2 and Mythos 5.2 are one model with different safeguards. Anthropic says Fable 5.2 causes fewer unnecessary blocks than 5.1.
  • In Anthropic's own test, Mythos 5.2 found 10.2% more known flaws than Opus 5.5. Direct access runs through a verification program.

Anthropic announced upgrades to Claude Fable and Claude Mythos, now both at version 5.2. The company says they are the same underlying model. Fable carries strong cybersecurity and biology safeguards, so any user can have it. Mythos is for organizations that apply through Anthropic's Cyber Verification and Life Sciences programs. Anthropic says Fable 5.2's safeguards are as strong as those of 5.1 but wrongly block less.

Anthropic notes that Fable 5.2's safeguards will still block most cybersecurity tasks. Its internal CVE Rediscovery test gives a model a project as it stood before a bug was fixed, then asks it to find that bug. The October 2026 run covered 1,829 published flaws. Mythos 5.2 found 10.2% more of them than Opus 5.5, and 14.4% more than Mythos 5.1. These are Anthropic's own results.

The model is the same, so the product security buyers choose is the safeguard layer and who gets approved. Direct Mythos 5.2 access goes through the Cyber Verification Program. Claude Enterprise customers can reach it indirectly through Claude Security, which scans code. Budget holders should ask who decides whether they qualify, and how Fable's limits affect their own security work.

A WebPulse Brief: a short report of an important event, written by the WebPulse Newsroom with AI assistance and checked against the reporting below. When there is more to explain, we follow up with a full story. How we use AI.

Reporting: Anthropic.