Anthropic has announced the public release of Fable 5, its most advanced artificial intelligence model to date, while simultaneously introducing usage restrictions due to cybersecurity and safety concerns.
The company said Fable 5 demonstrates significant improvements in capabilities, including advanced software coding and debugging, the ability to handle complex research queries, and enhanced performance in image analysis. Despite its technological progress, Anthropic emphasised that the model will be made available to the general public under controlled conditions to mitigate potential misuse risks.
In addition, Anthropic introduced an unrestricted variant, Claude Mythos 5, which will be accessible only to vetted corporations, research institutions, and select participants involved in the Glasswing Project—a cybersecurity-focused initiative designed to strengthen software systems and protect digital infrastructure using artificial intelligence tools.
The company noted that Claude Mythos 5 will not be released to the public due to its highly advanced capabilities, particularly its ability to identify and potentially exploit software vulnerabilities, which could present serious security risks if used maliciously.
The announcement comes shortly after US President Donald Trump signed an executive order enabling voluntary government oversight of advanced AI systems, particularly those with cybersecurity implications, highlighting growing regulatory attention on frontier AI development.
