'Claude Mythos': Anthropic confirms testing its most advanced model yet
Anthropic is testing its most advanced AI model yet, called Claude Mythos, which is expected to be a major leap forward from its predecessors but also poses significant cybersecurity risks. The company plans to release the model initially to cyber defenders to help them fortify codebases against potential AI-driven exploits.
Anthropic is testing its most advanced AI model yet. The model, called Claude Mythos, was revealed in a data leak. The leak exposed nearly 3,000 unpublished assets, including draft blog posts and images. Claude Mythos features a new tier called Capybara, which is bigger and better than previous top models. The model is expected to perform significantly better on reasoning, coding, and cybersecurity tests. However, it also poses significant cybersecurity risks and could trigger a new wave of cyberattacks. Anthropic plans to release the model initially to cyber defenders to help them prepare for potential AI-driven exploits.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.