Anthropic has introduced Claude Fable 5, a new Mythos-class artificial intelligence model built for safe general use. The company announced the launch on Tuesday and described it as its most capable public model to date.
According to Anthropic, Claude Fable 5 delivers stronger performance across several professional tasks. The model performs well in software engineering, knowledge work, computer vision, and scientific research. It also handles longer and more complex assignments more effectively than previous Claude models.
Anthropic stated that Claude Fable 5 can work independently for longer periods. The model can manage advanced workflows while maintaining reliable performance.
Key highlights include:
- Enhanced software engineering capabilities
- Improved scientific research support
- Stronger computer vision performance
- Better handling of complex tasks
- Extended autonomous operation
The company has implemented conservative safeguards to reduce misuse risks. These measures focus particularly on cybersecurity-related activities.
Certain requests will automatically route to Opus 4.8 instead of Claude Fable 5. Anthropic expects these safeguards to activate in less than five percent of average sessions.
Most cybersecurity, chemistry, and biology questions will receive responses from Opus 4.8. This approach aims to maintain responsible AI deployment while supporting legitimate use cases.
Anthropic also released Claude Mythos 5 to selected cyber defenders and infrastructure providers. The model uses the same foundation as Fable 5 but operates with fewer restrictions.
Mythos 5 is part of Project Glasswing and replaces the Claude Mythos Preview. Anthropic claims it offers the strongest cybersecurity capabilities available in any model today.
Both Claude Fable 5 and Mythos 5 cost $10 per million input tokens and $50 per million output tokens.
Claude Fable 5 remains available through Pro, Max, Team, and seat-based Enterprise plans until June 22. After that date, users must spend credits for access. Anthropic plans to restore subscription access once capacity reaches a stable level.
The announcement marks an important step in Anthropic's AI roadmap while balancing capability and responsible deployment.
Post a Comment