On Tuesday, Anthropic released Fable and Mythos 5.1, twinned versions of the company’s most advanced AI model. In addition to performance upgrades, the new Fable release includes changes meant to reduce token cost and false-positive restrictions from the model’s safeguards.
As with the previous Mythos model, Mythos 5.1 will only be available to registered Anthropic partners engaged in either cybersecurity or life sciences research. Fable 5.1, the unrestricted version, is available starting today on cloud platforms or through the Anthropic API.
One of the most significant changes is Anthropic’s previously reported embrace of zero data retention, allowing clients to run Anthropic models on their own infrastructure without data outflows. Previously unavailable for Fable due to security concerns, a high-privacy service (called Enterprise Frontier Safeguards) will now roll out to users in the fall. Notably, the system will still monitor for misuse by agents or human users, but clients will control how the monitoring takes place.
As part of the announcement, Anthropic assured customers that their data had not been inappropriately accessed. “Anthropic has never trained on enterprise data without explicit permission, and never will,” the announcement reads.
As is common for an Anthropic release, the new models set records in a range of benchmarks, including Terminal-Bench 4.0 (for CLI-based coding) and Humanity’s Last Exam (for general reasoning). Anthropic also released three novel scientific findings generated by the models before their release, including a custom GPU optimization and a high-resolution map of Venus assembled from existing photos.
As with previous releases, the models come with a detailed system card, which explains their capabilities in most straightforward terms. The system card rates Mythos as “low-risk” for concerns related to automated AI development — where the AI improves itself — which some see as a trigger for a loss of human control. It says “its ability to accelerate internal AI R&D progress is in line with current trends.”
In terms of general misbehavior, Mythos is slightly more prone to it than Opus, possibly as a result of its enhanced capabilities.
“Mythos 5.1 is a slight regression on overall misaligned behavior compared to Opus 5, and an improvement over Mythos 5 and Claude Sonnet 5,” the system card reads. “It cooperates with human misuse and accepts unverifiable claims of authorization somewhat more readily than Opus 5, but it is less likely to ignore explicit constraints, hallucinate inputs, or falsely claim to have completed tasks than previous models.”
