NewsBite analysis · AI, Cloud, Programming · New AI Model, Software Release
TestingCatalog
Anthropic launches Claude Fable 5.1 and Mythos 5.1
Anthropic has officially launched Claude Fable 5.1 and Claude Mythos 5.1, which offer improved safeguards and performance in AI-driven tasks like coding, knowledge work, and scientific research. Fable 5.1 is now generally available with a notable reduction in cache read costs by 75%, while Mythos 5.1 remains restricted to vetted users in cybersecurity and life sciences. These updates suggest not only cost efficiency but also advancements in accuracy and reliability of outputs.
Act nowConfidence: HighImmediate horizon
Published 2 Sept 2026, 00:55 UTC2 min read
Visit original article ↗Lead image — TestingCatalog
Summary
TL;DR
- Fable 5.1 features a 75% cost reduction in cache reads.
- Mythos 5.1 is available for vetted cyber defenders and life scientists.
- Anthropic introduced Enterprise Frontier Safeguards for data protection.
- Fable scored significantly higher in benchmark tests compared to earlier versions.
- The release includes an invisible watermark for compliance with the EU AI Act.
Interesting facts
- Fable now costs $0.25 per million tokens for cache reads.
- Mythos achieved a near 50% hit rate in designing protein binders.
- The Venus elevation map produced by Fable will be released under Creative Commons.