MEGA-Bench Task - Multilingual Game Info Parsing: leaderboard
Metric: Task Score (%). Source: huggingface.co. 44 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2 VL 72B | 88.4 |
| 2 | Gemini 2.5 Pro (Preview 03-25) | 85.7 |
| 3 | GPT-4o | 83 |
| 4 | Claude 3.5 Sonnet (20240620) | 83 |
| 5 | InternVL3-38B | 80.4 |
| 6 | GPT-4o Mini | 78.6 |
| 7 | InternVL2.5-78B | 77.7 |
| 8 | Qwen 2 VL 7B | 74.1 |
| 9 | InternVL3-8B | 73.2 |
| 10 | Claude 3.5 Sonnet (20241022) | 68.8 |
| 11 | InternVL3-78B | 67.9 |
| 12 | Gemma 3 12B (IT) | 67 |
| 13 | Gemma 3 4B (IT) | 60.7 |
| 14 | Gemma 3 27B (IT) | 58.9 |
| 15 | Llama 4 Scout Base | 54.5 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-multilingual-game-info-parsing · How It Works · Data refreshed daily, snapshot 2026-09-05.