MME-Safety - Harmful Rate: leaderboard
Metric: Harmful rate (%; share of final answers to 3,801 manually verified image-text pairs with harmful intent across nine risk categories and nine image/text stealth combinations, zero-shot; GPT-4o judge whose response contains harmful content; lower is better). Source: arxiv.org. Saturation forecast: Around December 2026. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 VL 235B A22B Instruct | 11.31 |
| 2 | Claude Sonnet 4 (Thinking) | 11.87 |
| 3 | Claude 3.7 Sonnet | 12.8 |
| 4 | Claude 3.7 Sonnet (Thinking) | 13.96 |
| 5 | O4 Mini | 15.32 |
| 6 | Qwen 3 VL 235B A22B (Thinking) | 15.34 |
| 7 | Qwen 3 VL 32B (Thinking) | 17.5 |
| 8 | Qwen 3 VL 8B (Thinking) | 19.83 |
| 9 | Gemini 3 Pro | 34.42 |
Interactive version: theaggregate.ai/benchmark?slug=mme-safety-harmful-rate · How It Works · Data refreshed daily, snapshot 2026-09-26.