Multimodal Reranking on MMEB VisDoc v2
87.1Overall ScoreQwen3-VL-8B-reranker
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Qwen3-VL-8B-rerankerModel Scale=8B, Fine-tuning Recipe=our training recipe2026.06 | 87.1 | 94.8 | 68.2 | 94 | 76.4 | |
| Qwen3-VL-8B-miniRerankerModel Scale=8B, Fine-tuning Recipe=our training recipe, Reranker Type=miniReranker2026.06 | 86.7 | 94.1 | 68.6 | 93.4 | 76.3 | |
| Qwen3-VL-4B-rerankerModel Scale=4B, Fine-tuning Recipe=our training recipe2026.06 | 86.5 | 94.3 | 66.8 | 93.9 | 75.9 | |
| Qwen3-VL-4B-miniRerankerModel Scale=4B, Fine-tuning Recipe=our training recipe, Reranker Type=miniReranker2026.06 | 86.2 | 93.8 | 68.1 | 92.7 | 75.3 | |
| Qwen3-VL-Reranker-8BModel Scale=8B2026.06 | 85.8 | — | — | — | — | |
| Qwen3-VL-2B-rerankerModel Scale=2B, Fine-tuning Recipe=our training recipe2026.06 | 85.7 | 93.6 | 66.7 | 92.1 | 75.5 | |
| Qwen3-VL-2B-miniRerankerModel Scale=2B, Fine-tuning Recipe=our training recipe, Reranker Type=miniReranker2026.06 | 84.9 | 93.3 | 64.3 | 92.1 | 73.8 | |
| Qwen3-VL-Reranker-2BModel Scale=2B2026.06 | 83.2 | — | — | — | — | |
| Qwen3-VL-2B-InstructModel Scale=2B2026.06 | 81.2 | 87.2 | 71.1 | 83.8 | 72.4 | |
| GME (7B)Model Scale=7B2026.06 | 75.2 | 89.4 | 55.6 | 85 | 44.4 | |
| GME (2B)Model Scale=2B2026.06 | 72.7 | 86.1 | 54 | 82.5 | 43.1 | |
| ColPali v1.3 (3B)Model Scale=3B2026.06 | 71 | 83.6 | 52 | 81.1 | 43.1 | |
| LamRA-Qwen2.5 (7B)Model Scale=7B2026.06 | 50.2 | 56.3 | 33.3 | 58.2 | 40.1 | |
| VLM2Vec-Qwen2VL (7B)Model Scale=7B2026.06 | 46.4 | 56.9 | 9.4 | 59.1 | 38.1 | |
| VLM2Vec (2B)Model Scale=2B2026.06 | 41.6 | 49.8 | 13.5 | 51.8 | 33.5 |