ArabicDialectSafety - Unsafe Response Rate: leaderboard
Metric: Unsafe response rate (%; share of 500 stratified harmful dialectal Arabic test prompts whose response a GPT-5.2 judge labels unsafe under a binary rubric; lower is better). Source: arxiv.org. Saturation forecast: Estimated already saturated. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.6 Plus | 0.2 |
| 2 | Claude Sonnet 4.6 | 0.6 |
| 3 | Claude 3 Haiku | 1.6 |
| 4 | Gemini 2.5 Flash | 2 |
| 5 | GPT-4o Mini | 4.6 |
Interactive version: theaggregate.ai/benchmark?slug=arabicdialectsafety-unsafe-response-rate · How It Works · Data refreshed daily, snapshot 2026-09-26.