Anthropic Lets Public Peek At Its Mythos Model

Anthropic, a San Francisco-based artificial intelligence company, is allowing the public to access a version of its powerful Mythos model for the first time. On Tuesday (June 9), the company launched Claude Fable 5, a model designed for general use with safety measures in place. According to TechCrunch, Fable 5 excels in software engineering, knowledge work, and vision tasks, but restricts responses in high-risk areas like cybersecurity, biology, and chemistry.

Initially, Mythos was available only to select government and corporate clients due to security concerns. However, Anthropic has now expanded access to hundreds of organizations across 15 countries. The public can access Fable 5 through Anthropic’s Claude API and Enterprise plans. Until June 22, the model is available at no extra cost on certain subscription plans, after which usage credits will be required.

The model has demonstrated impressive capabilities, with The Decoder reporting that Fable 5 outperforms other models in complex tasks, such as software engineering and financial analysis. It also boasts high efficiency in processing tokens. Despite its strengths, the model is priced at $10 per million input tokens and $50 per million output tokens, which is double the cost of the previous Claude Opus 4.8 model.

Anthropic has emphasized the importance of safety, implementing AI classifiers to redirect potentially dangerous requests to the Claude Opus 4.8 model. This safety measure affects less than 5% of sessions. The company plans to continue refining the model's safeguards and expand access gradually.

As Anthropic prepares for a potential public market entry, it remains committed to ensuring its models are used responsibly. With the launch of Claude Fable 5, the company aims to balance accessibility with safety, setting a precedent for future AI developments.