MEGA-Bench Task - Red Teaming Visual Misleading — leaderboard
Metric: Task Score (%). Source: huggingface.co. 44 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2 VL 72B | 90.5 |
| 2 | Qwen 2 VL 7B | 90.5 |
| 3 | Gemini 2.5 Pro | 90 |
| 4 | Gemma 3 27B (IT) | 90 |
| 5 | MiniCPM-V-2.6 | 89.5 |
| 6 | Pixtral-12B | 89.5 |
| 7 | Claude 3.5 Sonnet (20241022) | 88.4 |
| 8 | Gemini 1.5 Pro (002) | 88.4 |
| 9 | InternVL3-38B | 87.9 |
| 10 | Gemini 1.5 Flash (002) | 87.9 |
| 11 | Claude 3.5 Sonnet (20240620) | 87.9 |
| 12 | Gemini 2.0 Flash (Preview) | 87.9 |
| 13 | InternVL2-8B | 86.8 |
| 14 | Qwen 2 VL 2B | 86.3 |
| 15 | Gemma 3 4B (IT) | 85.8 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-red-teaming-visual-misleading · How the rankings work · Data refreshed daily, snapshot 2026-07-22.