KIDBench (No Cue) - School Conduct and Integrity: leaderboard
Metric: Category score (1-5): mean of the five child-safety dimensions, each rated 1-5 by a DeepSeek-V4-Pro judge with the KIDBench rubric and category-specific rules, temperature 0 for responder and judge, over the 50 School Conduct and Integrity prompts of KIDBench for children aged 7 to 11, single-turn, no-cue prompts (no sign that the user is a child); higher is better. Source: arxiv.org. Saturation forecast: Around June 2028. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 Mini | 3.75 |
| 2 | Qwen 3.6 27B | 3.51 |
| 3 | DeepSeek V4 Flash | 3.46 |
| 4 | Qwen 3.5 4B | 3.13 |
| 5 | Claude Haiku 4.5 | 3.06 |
| 6 | Gemma 3 12B | 2.96 |
| 7 | Qwen 3 8B | 2.9 |
| 8 | Gemini 3.1 Flash Lite | 2.77 |
| 9 | Gemma 3 4B | 2.75 |
| 10 | Gemma 4 31B (IT) | 2.64 |
| 11 | Llama 3.3 70B Instruct | 2.44 |
| 12 | Llama 3.1 8B Instruct | 2.33 |
Interactive version: theaggregate.ai/benchmark?slug=kidbench-no-cue-school-conduct-and-integrity · How It Works · Data refreshed daily, snapshot 2026-10-07.