Anthropic has rolled out two new frontier AI models, Claude Fable 5.1 and Claude Mythos 5.1, positioning them as the most capable systems yet for software development, enterprise knowledge work, and scientific research.

While both models share identical underlying architecture, they differ in one key respect: the strength of their safeguards. Fable 5.1 is generally available to all users, whereas Mythos 5.1 is restricted to vetted professionals through Anthropic’s trusted access programs, with permissions tailored specifically for cybersecurity defenders and life sciences researchers.

The release addresses three recurring pain points enterprise customers have raised about earlier Claude models: cost, data privacy, and overly aggressive content safeguards.

On pricing, Anthropic says Fable 5.1 will run roughly 25 percent cheaper than Fable 5 for typical workloads, and up to 45 percent cheaper for heavily agentic tasks, largely due to a 75 percent price cut on cached input reads, now $0.25 per million tokens. Input and output token pricing otherwise remains unchanged at $10 and $50 per million tokens, respectively.

Claude Fable and Mythos 5.1

Data handling gets a significant overhaul through Anthropic’s new Enterprise Frontier Safeguards (EFS) framework, which lets organizations store their data entirely within their own cloud infrastructure rather than on Anthropic’s servers, delivering zero-data-retention-level privacy while still catching adversarial misuse.

EFS will roll out in phases starting this fall across Claude Code, Claude Enterprise, the Claude Platform, and major cloud partners including AWS, Google Cloud, and Microsoft Azure. Until then, eligible enterprise customers can already use Fable 5.1 under a zero-retention policy.

On the safeguards front, Anthropic reports a 60 percent drop in false-positive interventions for cybersecurity-related queries, partly because Fable 5.1 is now permitted to identify software vulnerabilities for defensive purposes, though it still cannot generate exploits. Higher-risk dual-use tasks such as penetration testing and binary vulnerability scanning are automatically redirected to Anthropic’s Opus-class models instead.

Anthropic’s internal testing shows Fable 5.1 outperforming its predecessor across coding, agentic reasoning, and knowledge-work benchmarks, including Terminal-Bench 4.0, GDPval-AA v2, and Humanity’s Last Exam.

Early enterprise partners including Jane Street, Cognition, MongoDB, Datadog, and Millennium reported the model solving problems that had stumped engineering teams for years, including a rare production crash that had eluded human debuggers for nearly five years.

Mythos 5.1’s research capabilities extend into scientific domains. In molecular biology testing, the model designed high-affinity protein binders that outperformed leading entries in Adaptyv Bio’s design competitions, achieving nearly a 50 percent hit rate against a typical industry benchmark of 10 to 15 percent.

Separately, Fable 5.1 built a new high-resolution elevation map of Venus using decades-old NASA Magellan radar data, ahead of upcoming ESA and NASA missions.

Mythos 5.1’s advanced biology and cybersecurity capabilities remain gated behind Anthropic’s Cyber Verification Program and Life Sciences Verification Program, developed in coordination with the US government, with plans to broaden eligibility over time.

Anthropic has also confirmed compliance with the EU AI Act’s Code of Practice, embedding invisible watermarking into outputs from models released after August 2, 2026, alongside a detection API now in private preview for regulators and researchers. Fable 5.1 is available immediately on Claude’s platforms and via the API under the identifier claude-fable-5-1.

Prevent incidents due to slow investigations. Power your Tier 1 with threat intelligence from 15K SOCs: Integrate TI Lookup in your SOC

The post Anthropic Launches Claude Fable and Mythos 5.1 for Advanced Coding and Research appeared first on Cyber Security News.