Video Understanding on VideoMME (Temporal, Action, Object, Spatial Breakdown)
60.8TemporalGemini-3.1-Pro
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Gemini-3.1-Pro2026.04 | 60.8 | 61.9 | 60 | 51.3 | 59.8 | |
| VTHM-MoE2026.04 | 52.3 | 61.1 | 61.5 | 65 | 61.4 | |
| Qwen3-VL-235B2026.04 | 36.7 | 59.6 | 55.8 | 37.5 | 53.7 | |
| BaselineCoT=true, SFT=true2026.04 | 35.8 | 56.4 | 63 | 57 | 53.9 | |
| Internvl3.5-VL-241B2026.04 | 34.2 | 57.1 | 52.1 | 37.3 | 51.5 | |
| BaselineCoT=true2026.04 | 32.7 | 47 | 48.4 | 44.7 | 44.1 | |
| Qwen3-VL-30B2026.04 | 32.2 | 56.4 | 53.2 | 45.2 | 50.4 | |
| Baseline2026.04 | 23.4 | 30.5 | 37.5 | 31.2 | 31 | |
| BaselineSFT=true2026.04 | -33.8 | 46.5 | 52.8 | 36.3 | 51.6 |