Anthropic's Claude Fable 5.1 and Mythos 5.1 arrive with a 75% cost reduction for Fable cache reads
Anthropic unveiled Claude Fable 5.1 (GA) and Claude Mythos 5.1 (restricted)—the same model with differing safeguards—and announced a 75 % cost reduction for cached‑context reads, enabling far cheaper persistent agents. The models achieve strong benchmark gains, e.g., 52.6 % on Terminal‑Bench‑Science 0.1 (vs 24.7 % for Fable 5) and 55.8 % on Terminal‑Bench 4.0 (vs 42.0 % for Fable 5). Enterprise Frontier Safeguards (EFS) give organizations control over monitoring data, addressing security concerns from recent unauthorized‑action incidents.
⚡ Key Takeaways
- Cost of cached context reads reduced by 75 % for Fable 5.1.
- New security layer called Enterprise Frontier Safeguards (EFS) for on‑prem monitoring data retention.
- Benchmark: 52.6 % on Terminal‑Bench‑Science 0.1, 55.8 % on Terminal‑Bench 4.0, 73.4 % on CursorBench 3.2.0.
- Model supports sustained problem‑solving agents, demonstrated by a 38‑hour unattended ML run that launched experiments and returned findings.
- Mythos 5.1 operates with more permissive safeguards and reaches 60.9 % on the same coding benchmark.
The cost cut and EFS architecture make long‑running, high‑impact agents financially viable and compliant for enterprises, accelerating adoption of autonomous AI workflows in regulated environments.
✅ Practical Steps
- Deploy Claude Fable 5.1 for production agents to benefit from the 75 % cached‑context cost reduction.
- Integrate Enterprise Frontier Safeguards (EFS) to keep monitoring data within your own infrastructure when using Mythos 5.1.
Want the full story? Read the original article.
Read on VentureBeat AI ↗