News Ars Technica - All content Jun 9
Anthropic says these topics are too dangerous to let its Fable 5 model talk about
By Kyle Orland
86 score
AI Analysis
Anthropic publicly launched Claude Fable 5, its first widely available Mythos-class model, claiming it surpasses prior Opus frontier models. The release ships with guardrails that route queries on cybersecurity, biology, and chemistry to weaker models to avoid uplifting malicious actors, while the full Mythos 5 stays restricted to vetted cyberdefenders via Project Glasswing.
Anthropic Tuesday publicly released Claude Fable 5, its first "Mythos-class" model that it says surpasses its previous frontier Opus models in overall capabilities. But the model's launch today comes with safeguards designed to prevent it from answering queries on topics like cybersecurity, biology, and chemistry, where the company has publicly worried about its potential impact to "uplift" malicious actors.
Anthropic says Fable 5 operates on the "same underlying model" as Mythos 5, which is com
Model ReleaseAI SafetyDangerous Capabilities
Source details
Anthropic Tuesday publicly released Claude Fable 5, its first "Mythos-class" model that it says surpasses its previous frontier Opus models in overall capabilities. But the model's launch today comes with safeguards designed to prevent it from answering queries on topics like cybersecurity, biology, and chemistry, where the company has publicly worried about its potential impact to "uplift" malicious actors.
Anthropic says Fable 5 operates on the "same underlying model" as Mythos 5, which is com
Model ReleaseAI SafetyDangerous Capabilities