Trust in Rob
AI SummaryCurated from 1 authoritative sources

Claude Mythos 5.1 is Anthropic’s restricted model for vetted cyberdefenders and life scientists, sharing weights with generally available Claude Fable 5.1.

Anthropic introduces Claude Mythos 5.1, its newest Mythos-class model and the company’s most capable system for cybersecurity and biology research. The 1 September 2026 update keeps access limited to a small set of vetted organisations through trusted access programmes, while Claude Fable 5.1 — the same underlying model with tighter safeguards — is generally available.

Claude Mythos 5.1 is offered to vetted cyberdefenders and life scientists. The Life Sciences Verification Program is launching as an invite-only beta with reduced biology safeguards for advanced researchers. The Cyber Verification Program, which already provides reduced cyber safeguards for defensive security work, will include Mythos models in the near future. Access is currently limited to a set of US organisations, with expansion planned. Claude Security now also runs on Mythos 5.1. Pricing starts at $10 per million input tokens and $50 per million output tokens.

Because Mythos 5.1 is highly capable in cybersecurity and biology, Anthropic says it could be used for good or harm. Fable 5.1 therefore ships with more precise safeguards: it can identify software vulnerabilities in source code, and biology safeguards intervene on benign requests 85% less often than those launched with Fable 5. Dual-use biology and chemistry questions still route to Opus models. Penetration testing, exploit generation, and binary-based vulnerability scanning remain blocked. Using Mythos 5.1 requires a 30-day data retention policy for safety monitoring by default.

Anthropic reports early scientific results from the 5.1 family, including Mythos 5.1 protein binders with a nearly 50% hit rate across 12 targets, GPU kernel speed-ups of up to 2.5 times on open-source biology models, and a Fable 5.1 high-resolution elevation map of Venus derived from NASA Magellan radar. Safety testing found stronger cyber capabilities than prior releases, still within Anthropic’s lower risk category. Export controls on Mythos 5 were lifted on 1 July 2026 after a June disruption.

Read more on Anthropic’s Claude Mythos page.

Original Sources

Disclaimer

This is an AI-summarized article from authoritative sources. If you want your content removed, please .

WEEKLY DIGEST

Expert Insights Delivered Directly to Your Inbox.

Join thousands of readers who trust Rob for curated tech insights, practical automation tips, and strategies to reclaim your time.

Anthropic Introduces Claude Mythos 5.1 for Cybersecurity and Biology | TrustInRob