Anthropic launches AI model similar to Mythos with reinforced cybersecurity systems

Artificial intelligence firm Anthropic has launched Claude Fable 5, a new general-purpose AI model that boasts significant improvements in programming, data analysis, and complex knowledge work. The company stated the model is its most powerful yet for paid and enterprise clients.

“

The launch of a model with such capability carries risks. Without safety measures, Fable 5’s capabilities in areas like cybersecurity could be misused to cause serious harm.

Anthropic acknowledged the potential risks associated with such a capable model, particularly in cybersecurity. To mitigate these risks, Claude Fable 5 incorporates safety measures. These safeguards may result in certain sensitive queries being rerouted to a more advanced model, Claude Opus 4.8. The company noted that these conservative safety settings might occasionally block harmless requests, affecting less than 5% of user sessions on average.

In parallel, Anthropic is expanding its Project Glasswing initiative, which provides access to its cybersecurity-focused AI model. The project will now include 150 partners across more than 15 countries, encompassing sectors like energy, water supply, health, and communications. This expansion follows an initial April rollout that included major tech and security firms such as Amazon, Apple, Google, Microsoft, and Nvidia.

“

To launch the model safely and quickly, we have adjusted these safety measures with a conservative criterion; they will sometimes block harmless requests, although they are triggered, on average, in less than 5% of sessions.

The release of these new models comes shortly after Anthropic confidentially filed for an initial public offering with the U.S. Securities and Exchange Commission. This move positions Anthropic as a competitor to OpenAI, which is also reportedly preparing for a significant IPO.

“

Significant improvements in programming, data analysis, research, and complex knowledge work.