Anthropic Releases Its Most Powerful Claude Fable 5 Model to the Public

Anthropic has introduced its most advanced artificial intelligence model to the general public for the first time, though this process is being carried out with strict security measures. On Tuesday, the company launched the first version of its Mythos model intended for public use — the Claude Fable 5 model. According to Anthropic, Fable 5 demonstrates high performance in software engineering, intellectual labor, and visual analysis, but it comes with strict security restrictions. Techcrunch.com reports .

In high-risk areas such as cybersecurity, biology, and chemistry, the model blocks responses to queries and automatically switches to the Claude Opus 4.8 model. The Mythos model, introduced as a preview in April, was limited to a few partners due to cybersecurity concerns. Now, this technology is open to everyone through Anthropic Claude API and Enterprise plans. Additionally, the Fable 5 model will be available at no extra cost to Pro, Max, and Team subscribers until June 22.

Anthropic announced that it tested this model against various "jailbreak" attempts before its release. The company emphasizes that no universal vulnerabilities were found during more than 1000 hours of internal and external testing. Nevertheless, to protect against new types of attacks, Anthropic has implemented a policy of retaining all traffic data for 30 days. This data is not used for model training but serves solely for security purposes.

The release of the Fable 5 model coincides with Anthropic preparing to enter the mass market alongside companies like OpenAI and SpaceX, led by Elon Musk. Warning that artificial intelligence systems may reach the level of self-improvement without human intervention (RSI), the company is calling on global laboratories to coordinate security measures.