Multimodal Video Understanding on OmniParsingBench Text-Rich Video Module
66.78Overall ScoreLogics-Parsing-Omni
Evaluation results
| Method | Links | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Logics-Parsing-Omni2026.03 | 66.78 | 30.43 | 53.14 | 0.2 | 54.52 | 70.92 | 57.95 | 71.39 | 58.59 | 65.16 | 77.17 | 79.42 | 89.41 | 79.03 | |
| Gemini-3-ProModel Version=Pro2026.03 | 61 | 22.8 | 52.6 | 0.15 | 53.47 | 57.92 | 26.14 | 39.04 | 14.45 | 41.8 | 66.3 | 79.95 | 90.54 | 68.53 | |
| Qwen3-Omni-30B-A3BParameters=30B2026.03 | 25.06 | 3.5 | 21.69 | 1.21 | 8.4 | 8.04 | 3.98 | 6.68 | 1.56 | 7.79 | 9.51 | 11.08 | 76.5 | 41.72 |