Anthropic Announces New AI Models Claude Fable 5.1 and Mythos 5.1
Responding to customer criticism, Anthropic has unveiled new models offering lower costs and improved performance.
Artificial intelligence company Anthropic has introduced the Fable 5.1 and Mythos 5.1 models, responding to criticisms regarding pricing, data retention, and security measures.
New Models and Cost Advantages
Taking customer criticisms into account, Anthropic announced its AI models that improve upon pricing, data retention policies, and overly strict security measures.
The Claude Fable 5.1 model offers stronger performance compared to its predecessor while providing overall cost reductions of 25 percent and up to 45 percent for complex tasks.
Expert and Industry Opinions
Dan Shipper, CEO of Every, stated that the Fable 5.1 model is the most powerful coding model they have used to date and communicates just like a normal human.
Aaron Levie, CEO of Box, noted that the new model successfully captured data nuances and ambiguities that the previous version missed in the same test.
Security and Privacy Improvements
Fable 5.1 has been made available to users with more nuanced security measures that are less likely to block basic biology questions.
The company will launch the Enterprise Frontier Safeguards system in the fall, ensuring complete privacy by storing data directly on the customer's cloud servers rather than its own.
Cybersecurity and Model Access
While Anthropic allows the Fable 5.1 model to be used for identifying software vulnerabilities, it continues to direct certain tasks such as penetration testing and exploit generation to Opus models.
While Claude Fable 5.1 has become accessible across all platforms, the Mythos 5.1 model is currently only available to Project Glasswing participants.