MEGA-Bench Task - FunQA Unexpected Action Humor Video: leaderboard
Metric: Task Score (%). Source: huggingface.co. 44 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o | 55.3 |
| 2 | Gemini 2.5 Pro (Preview 03-25) | 43.3 |
| 3 | GPT-4o Mini | 40.7 |
| 4 | Qwen 2 VL 72B | 40 |
| 5 | Gemini 1.5 Pro (002) | 39.3 |
| 6 | Gemini 1.5 Flash (002) | 39.3 |
| 7 | Claude 3.5 Sonnet (20240620) | 37.3 |
| 8 | Gemma 3 27B (IT) | 36.7 |
| 9 | InternVL3-8B | 35.3 |
| 10 | InternVL3-38B | 35.3 |
| 11 | InternVL3-78B | 32.7 |
| 12 | InternVL2.5-78B | 32.7 |
| 13 | Llama 4 Scout Base | 32 |
| 14 | Claude 3.5 Sonnet (20241022) | 28.7 |
| 15 | Pixtral-12B | 28.7 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-funqa-unexpected-action-humor-video · How It Works · Data refreshed daily, snapshot 2026-09-05.