CC Capabilities

Capabilities / Vendor framing

Claude Fable 5.1 and Mythos 5.1: advanced reasoning with different access

Anthropic General AI capability score 76/100 confidence 0.85
Category
Vendor framing
Capability
Agentic coding and professional task execution
Observed
2026-09-01
Thesis section
Appendix III, sections one and two: model capability and vendor threshold evidence

Claim

Anthropic reports advances in extended coding, research and professional document work. Fable and Mythos share a base model but apply different safeguards, affecting which tasks each can perform.

Evidence: vendor reported ยท Reviewed 2026-10-24

Access: Fable 5.1 is generally available; Mythos 5.1 is restricted to verified organisations in trusted access programmes.

Oracle verdict

Safeguards and access conditions are part of usable capability. The restricted model should not be presented as generally available, and benchmark comparisons must account for interventions.

Why it matters

Reviewed against primary sources on 24 October 2026. Records vendor claims and access conditions; does not establish measured labour displacement. Score is editorial capability pressure, not a benchmark result.

Custom GPT Ask the Oracle