Anthropic launched Claude Fable 5 on Tuesday as its first publicly available Mythos-class model. The release includes safeguards that route sensitive queries on cybersecurity, biology and chemistry to an earlier model. Subscribers can access it through June 22 before usage credits apply.
Anthropic announced the launch of Claude Fable 5 alongside Claude Mythos 5. Fable 5 runs on the same underlying model as Mythos 5 but includes stricter safeguards. The company stated that Fable 5 refused all harmful single-turn requests on cyberattacks and exploit development during testing. The model offers gains in software engineering, document analysis and vision tasks. It beat Pokemon FireRed using a minimal vision-only setup, a feat earlier models could not achieve. Anthropic also reported that Fable 5 outperformed its prior Opus 4.8 model as well as GPT-5.5 and Gemini 3.1 Pro on internal benchmarks. Mythos 5, which carries fewer restrictions, is available only through Project Glasswing to a small group of trusted cybersecurity partners. Anthropic said it is expanding access in consultation with the US government and plans a new trusted program for life sciences organizations. API access to Fable 5 costs $10 per million input tokens and $50 per million output tokens. Regular subscribers can use the model at no extra cost until June 22, after which usage credits will be required.
