Back to AI updates
AnthropicPolicy

Improving Fable 5's biology safeguards

If you've had Fable 5 refuse ordinary health or biology questions, that should happen far less often now. It's still deliberately cautious on higher-risk topics like virology or toxicology, routing those to a more restricted model — worth knowing if your work touches those areas.

Read the original source 1 min read

What changed

Anthropic refined Fable 5's safety classifiers, cutting false-positive fallbacks on biology-related requests by roughly 85% so the model can help with everyday health, education and clinical questions it previously blocked — while continuing to restrict and route dual-use requests in virology, toxicology and molecular design to the less-capable Opus 5 model.

Why it matters to your work

If you've had Fable 5 refuse ordinary health or biology questions, that should happen far less often now. It's still deliberately cautious on higher-risk topics like virology or toxicology, routing those to a more restricted model — worth knowing if your work touches those areas.

What to try next

If Fable 5 has previously refused an ordinary health or biology question for your work, ask it again this week. Routine clinical and education questions should now go through.

Source & availability

This is a summary of a published provider announcement, with practical context from SCALA’s feed.

Plan, region, workspace and rollout conditions may affect access in Australia. Check the original source for current availability.

www.anthropic.com