Anthropic has rolled out Claude Fable 5, its first Mythos-class AI model for common users. This initiative marks a major strategic shift for the company. Anthropic had previously limited access to the Mythos model due to concerns about its advanced capabilities in cybersecurity and scientific research.
The company explained that Fable 5 is its most powerful publicly released AI model. The model can tackle complex reasoning, software engineering, visual analysis, and long-duration tasks while operating within strict safety boundaries.
In a blog post, Anthropic said, “Releasing a model this capable comes with risks. Without safeguards, Fable 5’s capabilities in areas like cybersecurity could be misused to cause serious damage.”
One of the main features of Claude Fable 5 is its safety framework. Requests from users for information on cybersecurity, biology, chemistry, and related subjects may be directed to a more secure model, Claude Opus 4.8.
According to Anthropic, these measures have been developed after thorough evaluations and security assessments done by the company. The firm thinks this would help provide broader public access without having to disclose all the risks posed by Mythos.
Claude Fable 5 is accessible through Anthropic’s subscription plans, including Pro, Max, Team, and Enterprise. Furthermore, developers can use Claude AI through the Claude API, Amazon Bedrock, Google Vertex AI, and Microsoft Foundry. It costs $10 per million input tokens and $50 per million output tokens and is considered one of Anthropic's top AI models.
The company had previously argued that unrestricted access to Mythos-class systems could create cybersecurity risks. After extensive testing and red-teaming, Anthropic concluded that Fable 5 could be released safely with additional controls in place.
The launch of Claude Fable 5 could represent an important step in the deployment of advanced AI apps. Anthropic appears to be taking more initiative in upgrading AI technology. It also aims to maintain safety and governance measures without limiting the usage of Fable 5.
Also read: Microsoft GitHub Breach Exposes AI Developers, 73 Repos Hit by Credential-Stealing Malware