Anthropic proposes an industry-wide scale for rating AI jailbreak severity
The Cyber Jailbreak Severity framework, developed with input from Amazon, Microsoft and Google, grades jailbreaks from CJS-0 to CJS-4 as Fable 5 returns with new safeguards.
The Cyber Jailbreak Severity framework, developed with input from Amazon, Microsoft and Google, grades jailbreaks from CJS-0 to CJS-4 as Fable 5 returns with new safeguards.
With Claude Fable 5 back for subscribers under time-limited access, here are five long-horizon projects that make better use of the model than ordinary chat.
After a suspension of nearly three weeks, Anthropic is restoring general access to Fable 5 while new cyber safety classifiers roll out and coding tasks go to Opus 4.8.
Anthropic says the jailbreak cited by US officials was neither universal nor novel, and that comparable capabilities already exist in rival models, while it complies with the directive.
Days after launch, Anthropic disabled its two most capable models worldwide to comply with an emergency US export control directive tied to a reported jailbreak.
After criticism from developers and researchers, Anthropic says Claude Fable 5 will now tell users when a request is blocked or rerouted to a less capable model.
Claude Fable 5 brings Anthropic's Mythos-class technology to business and premium users with safety fallbacks, while Mythos 5 stays limited to approved partners.