Anthropic has launched the latest versions of its advanced AI models, Fable 5.1 and Mythos 5.1. This update brings improvements not only in performance but also in reducing the token costs and easing previous overly cautious content restrictions. While Mythos 5.1 remains limited to selected partners in cybersecurity and life sciences research, the more open Fable 5.1 is now broadly accessible through cloud platforms and Anthropic’s API. A key innovation with this release is the introduction of zero data retention capabilities, allowing clients to operate the models on their own infrastructure without Anthropic retaining sensitive data.

The new Fable version supports high-privacy enterprise services under a program called Enterprise Frontier Safeguards, scheduled for rollout in the fall. This system preserves client control over data monitoring and misuse detection, addressing security concerns without compromising privacy. Anthropic emphasized its strict policies against unauthorized use of enterprise data, assuring customers that it has never trained models on their information without permission. Additionally, the models demonstrate strong benchmark performances, including improvements in coding and reasoning tasks, reaffirming their cutting-edge status.

Mythos 5.1 is classified as low risk regarding automated AI development, which refers to AI systems improving themselves. Despite slight regression in some behavior metrics versus the prior Opus 5 model, Mythos 5.1 strikes a balance between increased capabilities and risk management. The system card notes that it is somewhat more cooperative with misuse attempts but less likely to hallucinate or ignore constraints, reflecting ongoing refinement in AI safety and reliability. Anthropic continues to provide detailed system documentation to help users understand capabilities and limitations clearly.

Beyond performance, Anthropic highlighted three novel scientific outputs generated before public release, such as GPU optimization techniques and a detailed map of Venus from existing photos. These advances showcase the broader research benefits of their models. Overall, the Fable 5.1 upgrade signifies Anthropic’s commitment to creating more affordable, flexible, and secure AI tools for commercial and enterprise users, reinforcing its competitive position in the evolving AI landscape.