Anthropic’s Fable 5 is a powerful AI model built for complex reasoning, research, coding and advanced biology. Its biggest problem was over-cautious safety filtering which blocked even harmless questions about symptoms, laboratory reports and basic biology.
Anthropic has now rewritten and retrained the classifier that screens biology-related requests. The company says this has reduced unnecessary fallbacks by about 85%, allowing Fable 5 to answer many more health, educational and everyday science questions.
The controls have not been removed. Requests involving virology, toxicology, molecular design and advanced drug development may still be redirected because the same knowledge can support medical progress or enable biological harm. The aim is to distinguish legitimate learning from dangerous operational assistance.
For users, this means greater access to Fable 5’s expertise and the ability to understand complex information, analyse problems and produce high-quality work faster. Its power can create a “supercapability” by amplifying human knowledge and productivity—but growth will depend on trust, safety, responsible access and continued improvement of the controls.
POWER MUST GROW WITH RESPONSIBILITY.
Sanjay Sahay
Have a nice evening.