KORA Benchmark: Results as Published

KORA measures how AI systems behave with children and teenagers. The project names 26 child-safety risks and evaluates models on synthetic conversations meant to resemble questions children ask, not on sessions with real children. Scenarios are grouped by age band: Big Kids (7 to 9), Tweens (10 to 12), and Teens (13 and older). The public models leaderboard prints one score as a percentage, and a higher score means the model responded more appropriately to the scenarios.

Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.