Leading AI company Anthropic today unveiled Claude Fable 5, its first Mythos-tier model available to the public. The company states the model has passed safety evaluations and is suitable for general-purpose use. According to Anthropic, Fable 5 outperforms any previous public model from the company, delivering "exceptional performance" in areas such as software engineering, knowledge work, visual recognition, and scientific research. Notably, to prevent potential misuse, most queries involving cybersecurity, chemistry, and biology will be handled by the more conservative Opus 4.8 model rather than Fable 5 directly.
Alongside Fable 5, Anthropic also launched Mythos 5, deployed through the “Project Glasswing” initiative as an upgrade to the Claude Mythos Preview. Anthropic emphasizes that Mythos 5 offers the world's strongest cybersecurity capabilities, with access gradually expanding through the broader “Trusted Access Program.” Current Mythos Preview users — including Apple — can now use Mythos 5. Additionally, Anthropic released the latest Claude Opus 4.8 model, offering users at different levels more granular capability options.
