Anthropic on Tuesday released a restricted version of its powerful Mythos artificial intelligence system, a move that follows concerns that the original technology could be used by hackers to break into computer networks.
The new model, called Claude Fable 5, is twice as expensive as Anthropic's previous flagship system. The company said it includes additional guardrails designed to block responses related to cybersecurity, biology and other vulnerable areas.
A Delayed Release Over Security Fears
In April, Anthropic introduced Claude Mythos, a new AI system. But the company said it could not share Mythos with the public because the technology could become a powerful tool for hackers looking to penetrate computer networks. That warning alarmed executives and government officials who feared the system could trigger a new era of cybersecurity threats.
Now, with Claude Fable 5, Anthropic is offering a version that researchers believe is safe for widespread use. The company said Fable relies on older technology for some of its functions, which reduces its capabilities.
How Claude Fable 5 Differs
Because of the guardrails built into Claude Fable 5, hackers may find it difficult to use the system to attack computer networks. However, the same safeguards mean that businesses and cybersecurity experts may also struggle to use Fable to defend networks.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
When Claude users submit queries in areas that could be perceived as too risky, Anthropic said, those queries will be handled by Claude Opus 4.8. That model was released last month and was also designed to avoid the security risks of Mythos.
Anthropic's head of product management, Dianne Penn, explained the reasoning behind the approach. "That is the core of why we have been able to release Fable," she said.
Expanded Access to Mythos
Anthropic initially shared the full Mythos system with about 40 organizations that maintain critical computer infrastructure. Those groups used Mythos to patch security vulnerabilities before hackers could exploit them. This month, the company said it would expand access to Mythos to 150 additional organizations.
Anthropic is an AI safety company founded by former OpenAI employees. The company has developed a series of AI models under the Claude brand, with a focus on responsible deployment. The release of Claude Fable 5 represents a compromise between making advanced AI available and managing the risks of misuse.
The decision to restrict Mythos and offer a safer alternative highlights the ongoing tension between innovation and safety in the AI industry. Government regulators have been paying close attention to these developments, and the White House has issued executive orders related to AI security in recent years.
Related on Neura Market
- AI Tools Directory, Browse a curated list of artificial intelligence tools and platforms.
- Automation Marketplace, Discover automation solutions for business and security workflows.
- AI Models Category, Explore the latest AI model releases and comparisons.