SAN FRANCISCO: Artificial intelligence company Anthropic on Tuesday released Fable 5, the most powerful version of its AI technology made available to the general public, while maintaining restrictions designed to prevent misuse in sensitive areas such as cybersecurity and critical infrastructure.
Fable 5 is the first publicly accessible model from Anthropic’s advanced Mythos family, which was unveiled in April but initially faced limited deployment due to security concerns.
According to the company, the model is particularly effective at software coding, complex research tasks and image analysis. However, Anthropic said requests involving cybersecurity, biology and chemistry will be redirected to a less powerful model to reduce potential risks.
Alongside the public release, Anthropic is providing an unrestricted version, Claude Mythos 5, to selected organisations participating in its Project Glasswing programme. The initiative includes around 200 cybersecurity and research organisations across more than 15 countries.
Anthropic said it spent months testing the model’s safeguards, including extensive “red-teaming” exercises and bug bounty programmes aimed at identifying vulnerabilities. The company reported that no testers were able to completely bypass the restrictions.
The launch comes as competition intensifies in the AI industry. Both Anthropic and rival OpenAI recently announced plans that could pave the way for future stock market listings as they seek additional funding for increasingly expensive AI development.
Despite rapid revenue growth, Anthropic continues to invest heavily in computing infrastructure and recently signed a major data-centre agreement to support the growing demand for advanced AI systems.
