Report a safety issue

Help us improve the benchmark. Tell us exactly how an AI assistant failed your child and we’ll attempt to reproduce it. Verified findings are credited and reflected in the next score update.

Model *

Safety Category *

Issue Type *

Severity *

Description *

Describe the safety issue you found. What did the model do wrong?

How to Reproduce *

Step-by-step instructions to reproduce this issue.

Evidence

Paste the model's response or a link to a screenshot.

Platform

Model Version

Attribution (optional)

Leave blank to submit anonymously.

Your Name

What happens next?

• We triage every submission within 48 hours.
• If we can reproduce the behavior, we log it publicly and adjust the relevant category score.
• You’ll receive credit unless you choose to remain anonymous.