BREAKING
Anthropic Releases Claude Fable 5
0M
context tokens
0k
output tokens
AutomationBench Score
Fable 516.1
Opus 4.88.3
GPT-5.56.3
0$
per M input
0$
per M output
Strengths vs Concerns
Strengths
Multi-day autonomy
Refactoring and app builds
Self-verification
Concerns
High cost
Heavy token use
Slow, guardrails
Betting on Long-Horizon Autonomy
AI NEWS BLITZ
Anthropic has launched Claude Fable 5, its model built for long-horizon autonomy.