Anthropic · Model release
Claude Fable 5.1 and Mythos 5.1: the September refresh
Fable 5.1 lands at the top of the Artificial Analysis Intelligence Index at unchanged $10/$50 pricing — with a restricted-access Mythos 5.1 sibling for cyber and bio research.
What changed
- Debuted #1 on LiveBench (83.4, Max Effort) and the AA Intelligence Index (66) in launch week.
- Cache reads cut 75% to $0.25/MTok — roughly 25% savings on typical chat and ~45% on agentic workloads.
- Defaults to 'high' effort with adaptive thinking always on; knowledge cutoff June 2026. Mythos 5.1 (same underlying model, fewer cyber/bio safeguards) is invite-only for vetted US organisations.
Best fit
- Frontier agents
- Hardest reasoning work
- Fable 5 workloads at the same price
What to be careful about
- Leaderboard position is from launch-week boards; re-verify before committing spend.
- Cache-read savings only materialise when your harness actually reuses prompts — the default in Claude Code does.
Dataset snapshot: Claude Fable 5.1 is also represented in the checked comparison dataset. Prices and benchmark figures carry their own field dates.