Anthropic restricts Claude Fable 5 access, targeting Chinese AI labs

Anthropic has imposed restrictions on its latest AI model, Claude Fable 5, specifically targeting Chinese AI labs, creating new hurdles for their research. The model, designed for Mythos-class performance in software engineering and scientific research, includes built-in safeguards to prevent misuse. Restrictions rely on classifiers that flag queries related to cybersecurity, biology, chemistry, and frontier LLM development, including distillation attempts. When flagged, Fable 5 automatically downgrades the request to the older Claude Opus 4.8 model. The move has sparked debate about safety, innovation, and open research, with experts noting that Chinese AI developers may find it nearly impossible to use the model to accelerate their own development. Anthropic partially walked back enforcement after backlash from the global AI research community.