Anthropicのリスキーな「Mythos」モデル、ついに一般公開へ
Anthropicは初のMythosクラスモデルであるClaude Fable 5を公開した。高度なサイバー能力による安全上のリスクからこれまで公開を見送られていたが、新しい安全措置によって実現した。
anthropic released claude fable 5, the first broadly available model from its 'mythos' class, which the company previously withheld due to risks associated with its advanced cybersecurity capabilities. the release relies on new input-specific safeguards that restrict responses in sensitive domains like biology, with 95% of sessions processing natively without triggering fallbacks to the claude opus 4.8 model. pricing is set at $10 per million input tokens and $50 per million output tokens, double the rate of claude opus 4.8. concurrently, anthropic introduced claude mythos 5, an equivalent model with lifted safeguards available through a private trusted-access program for select enterprise partners.