SaLAD - Unsafe Queries: leaderboard

Metric: Accuracy rate (%; GPT-4o judge; 1,070 hazardous daily-life queries whose answer must name and explain the specific risk; a bare refusal is wrong). Source: arxiv.org. Saturation forecast: Around December 2026. 18 models tracked.

Top models

#ModelScore
1Claude 3.7 Sonnet57.2
2Gemini 2.5 Flash55.05
3GPT-4o53.83
4Llama 3.2 11B Vision33.46
5Qwen 2.5 VL 7B31.59
6MiniCPM-V-2.623.93

Interactive version: theaggregate.ai/benchmark?slug=salad-unsafe-queries · How It Works · Data refreshed daily, snapshot 2026-09-26.