MEGA-Bench Task - AV View Identification: leaderboard
Metric: Task Score (%). Source: huggingface.co. 44 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude 3.5 Sonnet (20241022) | 33.3 |
| 2 | GPT-4o | 32.2 |
| 3 | Gemini 1.5 Pro (002) | 27.8 |
| 4 | Claude 3.5 Sonnet (20240620) | 27.8 |
| 5 | Gemma 3 27B (IT) | 24.4 |
| 6 | Gemini 1.5 Flash (002) | 23.3 |
| 7 | Gemma 3 12B (IT) | 21.1 |
| 8 | InternVL2.5-78B | 21.1 |
| 9 | Gemini 2.5 Pro (Preview 03-25) | 21.1 |
| 10 | Llama 4 Scout Base | 21.1 |
| 11 | InternVL3-38B | 20 |
| 12 | InternVL3-78B | 18.9 |
| 13 | Gemma 3 4B (IT) | 16.7 |
| 14 | GPT-4o Mini | 14.4 |
| 15 | InternVL3-8B | 13.3 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-av-view-identification · How It Works · Data refreshed daily, snapshot 2026-09-05.