← Model benchmark
Anthropic · Claude Fable

Claude Fable 5

A high-capability Claude 5 model released with unusually conservative cyber and biology safeguards.

DOOM SCORE82.0out of 100
CURRENT ASSESSMENT · REVISION 1

Why this model scores 82.0

Strong agentic capability and general release elevate exposure, while stricter dangerous-domain safeguards materially lower residual control difficulty.

Capability91
Autonomy90
Deployment75
Misuse potential83
Control difficulty65
AUDIT TRAIL

Model score history

  1. R1
    Doom Score 82.0

    Initial source-backed model assessment

    11 Aug 2026