Temporal Grounding on SurgVidLM Out-of-domain (test)
69.96R@0.3ReWatch-R1-7B
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| ReWatch-R1-7BModel Class=Reasoning Agent2026.02 | 69.96 | 39.75 | 10.46 | 47.17 | |
| Video-RFTModel Class=Reasoning LMM2026.02 | 60.67 | 23.01 | 4.18 | 43.26 | |
| MedScope-7B-RLModel Class=Reasoning Agent, Training=RL2026.02 | 55.6 | 40.1 | 22.35 | 44.3 | |
| MedScope-7B-SFTModel Class=Reasoning Agent, Training=SFT2026.02 | 52.04 | 38.75 | 22.97 | 41.63 | |
| LongVT-7B-RFTModel Class=Reasoning Agent2026.02 | 51.13 | 37.66 | 16.73 | 39.74 | |
| VideoChat-R1-7BModel Class=Reasoning LMM2026.02 | 46.87 | 26.04 | 3.15 | 21.43 | |
| Gemini 3-Pro-PreviewModel Class=Proprietary LMM2026.02 | 45.19 | 31.59 | 22.8 | 37.35 | |
| Gemini-2.5-FlashModel Class=Proprietary LMM2026.02 | 43.51 | 33.89 | 23.64 | 33.95 | |
| Video-R1-7BModel Class=Reasoning LMM2026.02 | 41.18 | 23.07 | 9.42 | 31.45 | |
| Qwen2.5-VL-InstructModel Class=Open-Source LMM2026.02 | 27.61 | 21.94 | 15.31 | 19.15 | |
| InternVL3-8BModel Class=Open-Source LMM2026.02 | 20.92 | 14.23 | 10.88 | 18.55 | |
| GPT-4oModel Class=Proprietary LMM2026.02 | 15.9 | 9 | 5.65 | 13.35 | |
| VideoLLaMA3-7BModel Class=Open-Source LMM2026.02 | 12.67 | 10.08 | 5.21 | 9.37 |