MEGA-Bench Task - COCO OOD Global Image Retrieval By Query Property — leaderboard
Metric: Task Score (%). Source: huggingface.co. 44 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude 3.5 Sonnet (20240620) | 86 |
| 2 | GPT-4o | 85.8 |
| 3 | Gemini 2.5 Pro | 85.1 |
| 4 | Gemini 2.0 Flash (Preview) | 79.3 |
| 5 | Gemma 3 27B (IT) | 76.8 |
| 6 | Claude 3.5 Sonnet (20241022) | 74.5 |
| 7 | Qwen 2 VL 72B | 69.9 |
| 8 | Gemini 1.5 Flash (002) | 68.2 |
| 9 | Gemini 1.5 Pro (002) | 67.3 |
| 10 | InternVL3-38B | 66 |
| 11 | InternVL2.5-78B | 65.8 |
| 12 | InternVL3-14B | 64.9 |
| 13 | GPT-4o Mini | 64.9 |
| 14 | Gemma 3 12B (IT) | 64.8 |
| 15 | Pixtral-12B | 62.7 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-coco-ood-global-image-retrieval-by-query-property · How the rankings work · Data refreshed daily, snapshot 2026-07-22.