Anthropic Launches Claude Fable 5 and Mythos 5 with Different Safety Levels

Anthropic launched two new AI models: Claude Fable 5 for regular users and Claude Mythos 5 for cybersecurity defenders. They are the same model but with different safety settings. Fable 5 is the strongest Claude model available to the public, excelling in software engineering, knowledge work, vision, and science research. It can run longer on its own than any previous Claude model.

Fable 5 performs well in financial reasoning, document analysis, and chart understanding, and can extract exact numbers from complex scientific images. To prevent misuse, Anthropic added a classifier system that blocks requests related to cybersecurity, biochemistry, and distillation, switching to Claude Opus 4.8 when triggered.

Mythos 5 is built for high-sensitivity research and cybersecurity, with the world's strongest cybersecurity capabilities. It shines in drug design, molecular biology, and genomics. Internal protein design experts used Mythos 5 to speed up some processes by about 10 times, and out of 14 protein targets, 9 already have strong candidates.

Pricing is the same for both: $10 per million input tokens and $50 per million output tokens. Fable 5 is fully available now, while Mythos 5 is limited to Glasswing partners and a few biology researchers.