MVRB — leaderboard
Massive Visualized IR Benchmark leaderboard evaluating multimodal retrieval models across SR, CSR, SQA, and OVC components.
Metric: Overall Score. Source: huggingface.co. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | BGE-VL-Screenshot | 60.61 |
| 2 | GME | 48.14 |
| 3 | DSE | 45.21 |
| 4 | ColPali | 43.64 |
| 5 | MM-Embed | 34.48 |
| 6 | SigLIP | 33.34 |
| 7 | VLM2Vec | 32.19 |
| 8 | E5-V | 25.13 |
| 9 | CLIP | 23.75 |
| 10 | Uni-IR | 19.63 |
Interactive version: theaggregate.ai/benchmark?slug=mvrb · How the rankings work · Data refreshed daily, snapshot 2026-07-22.