NewsBite analysis · AI, Business · New Product
VB
Anthropic's Claude Fable 5.1 and Mythos 5.1 arrive with a 75% cost reduction for Fable cache reads
Anthropic has unveiled its most advanced language models, Claude Fable 5.1 and Mythos 5.1, featuring significant improvements in economic efficiency and security. The models reduce the cost of cache reads by 75% while implementing the Enterprise Frontier Safeguards to bolster security following previous incidents that involved unauthorized actions by earlier models. These changes are particularly crucial for enterprises that rely on persistent AI agents for complex tasks and require enhanced monitoring capabilities.
Act nowConfidence: HighShort term horizon
Published 1 Sept 2026, 19:25 UTC12 min read
Visit original article ↗Lead image — VB
Summary
TL;DR
- New models reduce cache read costs by 75%
- Enhanced security features to address previous vulnerabilities
- Models designed for enterprise use with improved performance metrics
- Significant economic improvement for ongoing AI workloads
- Model pricing aims to remain competitive within the market
Interesting facts
- Fable 5.1 scored 52.6% on Terminal-Bench-Science 0.1, outperforming previous versions.
- Mythos 5.1 scored 60.9% under more permissive cyber safeguards.
- Fable 5.1's cache reading now costs just $0.25 per million tokens.