MEGA-Bench Task - AV Human Multiview Counting — leaderboard
Metric: Task Score (%). Source: huggingface.co. 44 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Pixtral-12B | 33.3 |
| 2 | Qwen 2 VL 7B | 26.7 |
| 3 | Gemini 1.5 Flash (002) | 26.7 |
| 4 | Gemma 3 12B (IT) | 20 |
| 5 | Ovis2-8B | 20 |
| 6 | Aquila-VL-2B | 20 |
| 7 | Gemini 2.5 Pro | 13.3 |
| 8 | GPT-4o | 13.3 |
| 9 | InternVL3-8B | 13.3 |
| 10 | InternVL3-38B | 13.3 |
| 11 | InternVL3-14B | 13.3 |
| 12 | MiniCPM-V-2.6 | 13.3 |
| 13 | GPT-4o Mini | 13.3 |
| 14 | Llama 4 Scout Base | 13.3 |
| 15 | Gemma 3 27B (IT) | 6.7 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-av-human-multiview-counting · How the rankings work · Data refreshed daily, snapshot 2026-07-22.