tech
Anthropic Releases Claude Fable, a Version of Mythos, Days After Warning AI Is Becoming Too Dangerous
Anthropic is releasing Claude Fable 5, its first Mythos-class model available to the public. The model comes with guardrails that block responses in high-risk areas like cybersecurity and biology.
TL;DR
- Anthropic released Claude Fable 5, the first public version of its Mythos AI model.
- Fable 5 has safety limits; it blocks responses in high-risk areas like cybersecurity and biology, falling back to Claude Opus 4.8.
- The model was previously in preview and limited to select partners due to cybersecurity concerns.
- Fable 5 will be included in Pro, Max, Team, and Enterprise plans at no extra cost until June 22, after which usage credits will be required.
- Anthropic is also deploying a new version, Mythos 5, to approved organizations.
- The release comes as Anthropic prepares for public markets and reiterates warnings about AI's rapid advancement and potential for recursive self-improvement.
- Anthropic conducted extensive internal and external testing to prevent jailbreaks, but is implementing a mandatory 30-day traffic retention policy for safety.
- Early data shows Fable 5 sessions run entirely on its own responses at least 95% of the time.
- Third-party evaluations show Fable excels in complex analytical tasks, UI design, and game coding.
- Pricing for Fable 5 and Mythos 5 is double that of Opus 4.8, potentially acting as a deterrent for some.
- Some companies, like Rakuten, believe the advanced capabilities justify the higher cost for enabling autonomous operations.