KIDBench (No Cue): leaderboard
Metric: Total score (1-5): mean of the five child-safety dimensions (content safety, developmental appropriateness, socio-emotional support, moral guidance, boundary-setting), each rated 1-5 by a DeepSeek-V4-Pro judge with the KIDBench rubric and category-specific rules, temperature 0 for responder and judge, on the KIDBench child-safety prompts for children aged 7 to 11 (50 per category, 10 categories), single-turn, no-cue prompts (no sign that the user is a child); higher is better. Source: arxiv.org. Saturation forecast: Around 2032. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V4 Flash | 3.31 |
| 2 | Llama 3.3 70B Instruct | 3.14 |
| 3 | Qwen 3.6 27B | 3.14 |
| 4 | GPT-5 Mini | 3.06 |
| 5 | Claude Haiku 4.5 | 3.03 |
| 6 | Qwen 3.5 4B | 2.94 |
| 7 | Qwen 3 8B | 2.93 |
| 8 | Gemini 3.1 Flash Lite | 2.92 |
| 9 | Gemma 3 12B | 2.88 |
| 10 | Gemma 4 31B (IT) | 2.82 |
| 11 | Llama 3.1 8B Instruct | 2.81 |
| 12 | Gemma 3 4B | 2.74 |
Interactive version: theaggregate.ai/benchmark?slug=kidbench-no-cue · How It Works · Data refreshed daily, snapshot 2026-10-07.