Video Visual Question Answering on VSI-Bench
68.5ACC (MCA)GAMSI_S1+S2
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GAMSI_S1+S2SI Dataset=950K2026.05 | 68.5 | — | — | |
| GAMSI_S1SI Dataset=800K2026.05 | 66 | — | — | |
| Cambrian-S-7BSI Dataset=590K2026.05 | 63 | — | — | |
| SenseNova-SI-Qwen3-VL-8BSI Dataset=8M2026.05 | 62.9 | — | — | |
| SenseNova-SI-InternVL3-8BSI Dataset=800K2026.05 | 60.6 | — | — | |
| Qwen3-VL-8B-Instruct2026.05 | 56.9 | — | — | |
| Cambrian-S-3BSI Dataset=590K2026.05 | 55.9 | — | — | |
| VST-7B-SFTSI Dataset=4M2026.05 | 55.5 | — | — | |
| VST-3B-SFTSI Dataset=4M2026.05 | 51.7 | — | — | |
| Gemini-1.5 Protrained_with_MindCraft=false, Evaluation=Zero-Shot2026.05 | 49.7 | 44 | 45.4 | |
| Qwen3-VL-2B-Instruct2026.05 | 48.5 | — | — | |
| Spatial-MLLM-4BSI Dataset=135K2026.05 | 46.2 | — | — | |
| SpatialLadder-3BSI Dataset=260K2026.05 | 45.1 | — | — | |
| ViLaSR-7BSI Dataset=81K2026.05 | 44.9 | — | — | |
| LASAR (Ours)trained_with_MindCraft=true, Evaluation=Zero-Shot2026.05 | 43.5 | 52.9 | 48.9 | |
| InternVL3-8B2026.05 | 43.3 | — | — | |
| LASAR (IL+QA)trained_with_MindCraft=true, Evaluation=Zero-Shot2026.05 | 42.5 | 47.6 | 44.8 | |
| SpaceR-7BSI Dataset=151K2026.05 | 40.9 | — | — | |
| LASAR (IL)trained_with_MindCraft=false, Evaluation=Zero-Shot2026.05 | 40.1 | 35.5 | 37.8 | |
| Uni-NaVidtrained_with_MindCraft=true, Evaluation=Zero-Shot2026.05 | 37.1 | 42.5 | 39.7 | |
| NavCoTtrained_with_MindCraft=true, Evaluation=Zero-Shot2026.05 | 36.9 | 31.5 | 34.3 | |
| Uni-NaVidtrained_with_MindCraft=false, Evaluation=Zero-Shot2026.05 | 34.8 | 39.1 | 35.6 | |
| NavCoTtrained_with_MindCraft=false, Evaluation=Zero-Shot2026.05 | 34.7 | 29.6 | 32.1 | |
| GPT-4otrained_with_MindCraft=false, Evaluation=Zero-Shot2026.05 | 33.5 | 36.1 | 34 | |
| InternVL3-2B2026.05 | 33.5 | — | — | |
| Qwen2.5-VL-7B-Instruct2026.05 | 33.3 | — | — | |
| Bagel-7B-MoT2026.05 | 32.5 | — | — | |
| Qwen2.5-VL-3B-Instruct2026.05 | 27.5 | — | — |