After saying Mythos was too dangerous, Anthropic just launched a public version (www.techradar.com)

Anthropic has launched a public version of its powerful AI model, Fable 5, previously known as the Mythos-class, which the company had deemed too dangerous for general release due to potential cybersecurity and bioethical risks. This new model is now available on higher-tier subscription plans for Claude, representing a significant pivot in Anthropic's approach to deploying advanced AI technology. While the company has implemented extensive guardrails and classified monitoring systems to mitigate risks of misuse, such as detecting advanced cybersecurity activities, this release challenges the previous caution exhibited by Anthropic. Fable 5 is designed to tackle complex tasks over extended durations, making it suitable for in-depth research and multi-step projects rather than simply responding to single prompts. Although user feedback has been largely positive, concerns persist about accessibility and the disparities between the public and institutions that access more capable AI models. Initially included in subscription plans, Fable 5 will transition to a usage-credit model temporarily, aimed at managing demand, though Anthropic plans to restore it as a standard feature when capacity permits. This shift could reshape expectations about access to advanced AI technology, highlighting the ongoing debate about its implications for cybersecurity and general AI availability.