Video Understanding on VSI-Bench
49.5AccuracyAdaTooler-V-7B
Evaluation Results
| Method | Links | |
|---|---|---|
| AdaTooler-V-7BFrames=1282025.12 | 49.5 | |
| AdaTooler-V-7BFrames=642025.12 | 47.9 | |
| AdaTooler-V-7BFrames=322025.12 | 46.7 | |
| Gemini 1.5 Pro2025.12 | 45.4 | |
| VILASR2025.12 | 45.4 | |
| Gemini-2.0-Flash2026.05 | 45.4 | |
| InternVL3-8B2025.12 | 42.1 | |
| Qwen2.5-VL-7B + SAGESAGE=true2026.05 | 37.7 | |
| ViSS-R1-7B2026.05 | 37.3 | |
| Video-R12025.12 | 37.1 | |
| VideoRFT-7B2026.05 | 36.8 | |
| Qwen2.5-VL-7B + SFT + SAGESFT=true, SAGE=true2026.05 | 36.7 | |
| Qwen2.5-VL-7B-InstructFrames=1282025.12 | 34.8 | |
| Video-R1-7B2026.05 | 34.6 | |
| GPT-4o2025.12 | 34 | |
| GPT-4o2026.05 | 34 | |
| Qwen2.5-VL-7B2026.05 | 32.5 | |
| LLaVA-OneVision-7B2026.05 | 32.4 | |
| VILA-1.5-40B2026.05 | 31.2 | |
| Qwen2.5-VL-7B + SFTSFT=true2026.05 | 31.1 | |
| Qwen2.5-VL-7B-InstructFrames=642025.12 | 30.9 | |
| Qwen2.5-VL-7B-InstructFrames=322025.12 | 29.8 | |
| LongVA-7B2026.05 | 29.2 |