Key facts
- Anthropic AI has launched Fable 5, a Mythos-class large language model.
- Fable 5 includes safety circuit breakers for cybersecurity and bioweapons queries.
- Sensitive queries are rerouted to a less capable model.
- An unrestricted Mythos 5 tier is available to vetted government and infrastructure partners.
Anthropic AI has introduced Fable 5, a new large language model belonging to its Mythos class. This iteration is distinguished by integrated circuit breakers designed to manage queries concerning cybersecurity and bioweapons. These specific queries are automatically rerouted to a less advanced model, enhancing safety protocols. Concurrently, Anthropic is offering a distinct tier, Mythos 5, which lacks these restrictions. This unrestricted version is made available to select government and infrastructure partners who have undergone vetting processes. The company is rolling out a public version of its Mythos AI model with guard rails, barring its use in risky areas such as cybersecurity, after a preview earlier this year sent shock waves globally with its ability to find software flaws. The new Claude Fable 5 is the most powerful model Anthropic has ever made for wider use, the start-up said on Tuesday, touting its performance in software engineering and analytics.