Anthropic Unveils Claude Fable 5 AI Model with Enhanced Security Measures
Anthropic has released its most advanced artificial intelligence model, the Claude Fable 5, to the public. This model is designed to excel in software engineering, intellectual labor, and visual analysis. However, it is equipped with strict security measures to prevent misuse. In high-risk fields such as cybersecurity, biology, and chemistry, the model automatically switches to a safer version, the Claude Opus 4.8, to block potentially harmful responses. Initially previewed in April, the Mythos model was restricted to select partners due to cybersecurity concerns. Now, it is accessible to all users through Anthropic's Claude API and Enterprise plans, with no additional cost for Pro, Max, and Team subscribers until June 22. Anthropic has conducted extensive testing, including over 1000 hours of internal and external evaluations, to ensure the model's security. The company retains traffic data for 30 days for security purposes, although this data is not used for model training. 
