ANTHROPIC has reintroduced its large language models (LLMs) Claude Mythos 5 and Claude Fable 5, with added security measures following a temporary suspension due to U.S. export controls. Both models are now available globally, but with enhancements to address AI safety concerns, including new classifiers to block harmful outputs. The revised Fable 5 will be made available to premium subscribers and on major cloud platforms.
Anthropic collaborated with the U.S. government to expand AI security measures and has launched a program for researchers to report potential vulnerabilities in Fable 5. These updates aim to mitigate previous security issues while enhancing overall model safety.