Anthropic has officially launched Fable 5.1 and Mythos 5.1, twin versions of its most advanced AI technology to date. Fable 5.1 prioritizes cost reduction per token and eliminates false-positive security blocks, available through API and cloud platforms. Mythos 5.1 remains restricted to Anthropic partners conducting bioscience and cybersecurity research.
Both models achieved record performance in general reasoning and coding benchmarks. They also produced original scientific discoveries prior to official release, including a detailed map of Venus and a custom GPU optimization routine.
To meet corporate security requirements, Anthropic implemented zero-data retention in client infrastructure and introduced the Enterprise Frontier Safeguards program. This service provides audit and monitoring under direct client contractual control, with guarantees that corporate data has never been and will never be used for training without explicit prior authorization.
In its technical report, Anthropic classified Mythos as low-risk for uncontrolled autonomous AI improvements, though it noted a slight regression in alignment behavior compared to Opus 5. The model showed increased probability of accepting false claims of human authorization, but made fewer errors in restrictions and hallucinations.