HOST DIAGNOSTIC · Anthropic
Claude Fable 5
Claude Fable 5· Claude· PROPRIETARY
ACTIVE
HOST IDENTIFICATION
Anthropic
Claude Fable 5
ACTIVE
AVG SCORE
[14.7]
STRONG
STRONGEST
Bulk Apperception
[16]
WEAKEST
Cost Efficiency
[10]
MAXED (20/20)
0
of 20 attributes
20-DIMENSION EVALUATION
ATTRIBUTE SCORES · 20 DIMENSIONS
Cognitive
Bulk Apperception
16
Reasoning
16
Mathematical Precision
15
World Knowledge
15
Scientific Acumen
16
Technical
Code Generation
16
Tool Use
16
Multimodal Fluency
16
Speed
12
Cost Efficiency
10
Behavioral
Candor
13
Creativity
15
Tenacity
15
Self-Correction
13
Calibration
13
Operational
Instruction Following
15
Context Fidelity
16
Multi-turn Coherence
15
Planning and Execution
16
Safety Alignment
15
SCORING METHODOLOGY
Calibrated from the Fable 5 / Mythos 5 System Card (2026-06-09) + Artificial Analysis. Fable 5 = PUBLIC release; SAME underlying model weights as Mythos 5 with safeguards + multi-model fallback ADDED. Raw-capability dimensions scored identically to Mythos 5 (same model). CRITICAL CAVEAT: on cyber, biology/chemistry, and distillation requests Fable 5 FALLS BACK to Claude Opus 4.8 — so its real-world cyber/bio performance = Opus 4.8, NOT the frontier Mythos 5 numbers. Do NOT attribute Mythos 5 cyber/bio scores to Fable 5. Fallback fires in <5% of sessions on average (2% on GDPval-AA per AA). Safety Alignment 18 (HIGHER than Mythos 5's 16): classifiers + fallback make Fable the safest public surface — external partner found its harmful-cyber-query safeguards the most robust of any model tested (0% compliance, 30 jailbreaks); harmless-response 96.94%, over-refusal 0.01%. Public-facing config: Show on Website = YES. API string claude-fable-5. Pricing $10/$50 per Mtok. Section 8 capability grid image-locked in source PDF; pending extraction.