Multimodal Video Retrieval on MMEB Video Portion v2
56.8K700 ScoreEmbed-RL
Evaluation Results
| Method | Links | ||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Embed-RLModel Scale=4B2026.02 | 56.8 | 59.5 | 60.1 | 78.5 | 33 | 55.9 | 50.5 | 58.2 | 52.8 | 74.4 | 46.8 | 46.2 | 65.8 | 43.4 | 23.3 | 73.6 | 25 | 49.9 | |
| Embed-RLModel Scale=2B2026.02 | 55.8 | 56.7 | 56.7 | 79.3 | 36.7 | 50.8 | 47.1 | 53.9 | 53 | 74.8 | 45.3 | 45.7 | 67.2 | 43.6 | 23.5 | 70.7 | 26.4 | 50.9 | |
| UME-R1-7BModel Scale=7B2026.02 | 42.8 | 50.4 | 58.3 | 70 | 21.5 | 58.2 | 47.3 | 69.6 | 52.4 | 76 | 40 | 38.9 | 60.8 | 32.6 | 18.5 | 54.9 | 21.9 | 41.1 | |
| CAFe-7BModel Scale=7B2026.02 | 40.1 | 35.8 | 46.9 | 39.6 | 16.6 | 48.9 | 46 | 62.4 | 60 | 76 | 37.8 | 36.5 | 56.4 | 32 | 9.5 | 58.4 | 18.7 | 41.4 | |
| GME-7BModel Scale=7B2026.02 | 39.7 | 30.6 | 47.9 | 54.7 | 14.3 | 46.6 | 39.2 | 53.6 | 46.8 | 65.6 | 26.4 | 31.8 | 49.7 | 24.9 | 9.1 | 59.5 | 14 | 37.4 | |
| VLM2Vec-V2-2BModel Scale=2B2026.02 | 38 | 42.8 | 40.9 | 60 | 14.8 | 33.7 | 30.7 | 20.9 | 34 | 52.3 | 30.4 | 28.3 | 48.1 | 26.5 | 10.6 | 49.4 | 20.2 | 40.8 | |
| UME-R1-2BModel Scale=2B2026.02 | 35.8 | 44.1 | 54.4 | 67.2 | 20.1 | 49.9 | 41.7 | 59.9 | 45.4 | 57.8 | 32.4 | 34.3 | 55.4 | 29.9 | 12.7 | 57.5 | 20.4 | 41.2 | |
| VLM2Vec-7BModel Scale=7B2026.02 | 35.5 | 32.1 | 42.2 | 61.8 | 23.8 | 28.5 | 27.8 | 20.3 | 21.8 | 51.4 | 29.3 | 34.5 | 46.7 | 25.5 | 9 | 57.7 | 19.8 | 39.3 | |
| ColPali v1.3version=v1.32026.02 | 23.4 | 25.1 | 24.8 | 49.4 | 10.9 | 33.7 | 30.6 | 35.2 | 38.4 | 51.3 | 22.8 | 17.6 | 45.4 | 16.7 | 5.3 | 19.9 | 29 | 27.6 |