Video Hallucination Evaluation on VideoHallucer
82.36Overall ScoreDTR
Evaluation results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DTRBackbone=Qwen3-VL-4B2026.04 | 82.36 | 89.5 | 85.8 | 89 | — | — | — | — | — | 56 | 91.5 | — | — | — | — | |
| Qwen3-VL-4BBackbone=Qwen3-VL-4B2026.04 | 79.68 | 88 | 82.39 | 89 | — | — | — | — | — | 53 | 86 | — | — | — | — | |
| DINO-HEALBackbone=Qwen3-VL-4B2026.04 | 79.58 | 87.5 | 82.39 | 89 | — | — | — | — | — | 51.5 | 87.5 | — | — | — | — | |
| TCDBackbone=Qwen3-VL-4B2026.04 | 78.86 | 85.5 | 81.82 | 88 | — | — | — | — | — | 52.5 | 86.5 | — | — | — | — | |
| VCDBackbone=Qwen3-VL-4B2026.04 | 76.04 | 85 | 76.7 | 84 | — | — | — | — | — | 50 | 84.5 | — | — | — | — | |
| SELF-PEPBase Model=LLaVA-NeXT-Video-7B-DPO2026.05 | 74.4 | — | — | — | — | — | — | — | — | 11.5 | 15.5 | 52 | 37 | 19.3 | — | |
| LaSCDBase Model=LLaVA-NeXT-Video-7B-DPO2026.05 | 64.8 | — | — | — | — | — | — | — | — | 14.5 | 30 | 52 | 39.5 | 19.3 | — | |
| DeCoBase Model=LLaVA-NeXT-Video-7B-DPO2026.05 | 61.9 | — | — | — | — | — | — | — | — | 12.5 | 28.5 | 51 | 39 | 19.3 | — | |
| RegularBase Model=LLaVA-NeXT-Video-7B-DPO2026.05 | 60.9 | — | — | — | — | — | — | — | — | 14.5 | 30.5 | 51 | 38 | 27.8 | — | |
| MultiToPBackbone=Qwen3-VL-4B-Instruct2026.06 | 57.58 | — | — | — | — | — | -0.03 | 0.44 | — | — | — | — | — | — | — | |
| SEASONBackbone=QWEN2.5-VL-7B, Training-free=true2025.12 | 54.8 | 63.5 | — | 73.5 | 31.5 | 56 | — | — | 49.5 | — | — | — | — | — | — | |
| ArrowRLBackbone=QWEN2.5-VL-7B, Training-free=false2025.12 | 54.5 | 60.5 | — | 68 | 33.5 | 55.5 | — | — | 55 | — | — | — | — | — | — | |
| DINO-HEALBackbone=QWEN2.5-VL-7B, Training-free=true2025.12 | 53.3 | 61 | — | 71.5 | 32 | 55.5 | — | — | 46.5 | — | — | — | — | — | — | |
| QWEN2.5-VL-7BBackbone=N/A, Training-free=-2025.12 | 53.1 | 62 | — | 70.5 | 31.5 | 55 | — | — | 46.5 | — | — | — | — | — | — | |
| TCDBackbone=QWEN2.5-VL-7B, Training-free=true2025.12 | 52.8 | 61 | — | 70.5 | 31 | 55 | — | — | 46.5 | — | — | — | — | — | — | |
| SmartSightBackbone=Qwen3-VL-4B-Instruct2026.06 | 52.7 | — | — | — | — | — | -0.08 | 0.36 | — | — | — | — | — | — | — | |
| LaSCDBase Model=Video-LLaVA-7B2026.05 | 52.2 | — | — | — | — | — | — | — | — | 15 | 38.5 | 41 | 27 | 18.8 | — | |
| Video-ToC2026.04 | 51.9 | — | — | — | — | — | — | — | — | 20 | — | 66 | 74 | 45 | 54.5 | |
| LLaVA-Video-7BBackbone=N/A, Training-free=-2025.12 | 51.4 | 60 | — | 66.5 | 16.5 | 52.5 | — | — | 61.5 | — | — | — | — | — | — | |
| SEASONBackbone=LLaVA-Video-7B, Training-free=true2025.12 | 51.4 | 60.5 | — | 68 | 18 | 48.5 | — | — | 62 | — | — | — | — | — | — | |
| DINO-HEALBackbone=LLaVA-Video-7B, Training-free=true2025.12 | 51.3 | 59.5 | — | 66.5 | 17 | 52.5 | — | — | 61 | — | — | — | — | — | — | |
| TPOBackbone=LLaVA-Video-7B, Training-free=false2025.12 | 51.3 | 60 | — | 68.5 | 16 | 52.5 | — | — | 59.5 | — | — | — | — | — | — | |
| RRPOBackbone=LLaVA-Video-7B, Training-free=false2025.12 | 51.3 | 59 | — | 67 | 21 | 51.5 | — | — | 58 | — | — | — | — | — | — | |
| DINO-HEALBackbone=Qwen3-VL-4B-Instruct2026.06 | 50.57 | — | — | — | — | — | -0.33 | 0.16 | — | — | — | — | — | — | — | |
| BaselineBackbone=Qwen2.5-VL-7B2026.04 | 50.5 | — | — | — | — | — | — | — | — | 25.5 | — | 61.5 | 67.5 | 44 | 54 | |
| TCDBackbone=Qwen3-VL-4B-Instruct2026.06 | 50.47 | — | — | — | — | — | -0.11 | 0.3 | — | — | — | — | — | — | — | |
| TCDBackbone=LLaVA-Video-7B, Training-free=true2025.12 | 50.1 | 58 | — | 65.5 | 14.5 | 51.5 | — | — | 61 | — | — | — | — | — | — | |
| TA-AEBackbone=Qwen3-VL-4B-Instruct2026.06 | 49.41 | — | — | — | — | — | -0.1 | 0.32 | — | — | — | — | — | — | — | |
| SEASONBackbone=LLaVA-OV-7B, Training-free=true2025.12 | 48.5 | 63 | — | 56.5 | 19.5 | 48 | — | — | 55.5 | — | — | — | — | — | — | |
| TCDBackbone=LLaVA-OV-7B, Training-free=true2025.12 | 48.1 | 59.5 | — | 56 | 17.5 | 54 | — | — | 53.5 | — | — | — | — | — | — | |
| Qwen3-VL-4B-InstructBackbone=Qwen3-VL-4B-Instruct2026.06 | 47.69 | — | — | — | — | — | -0.09 | 0.34 | — | — | — | — | — | — | — | |
| DINO-HEALBackbone=LLaVA-OV-7B, Training-free=true2025.12 | 46.9 | 57 | — | 56.5 | 15.5 | 52 | — | — | 53.5 | — | — | — | — | — | — | |
| DTRBase Model=InternVL3.5-14B2026.04 | 46.84 | — | 68.18 | — | — | — | — | — | — | 11.5 | 48 | 50.5 | 56 | — | — | |
| TCDBase Model=InternVL3.5-14B2026.04 | 46.65 | — | 64.77 | — | — | — | — | — | — | 14 | 48 | 51 | 55.5 | — | — | |
| LLaVA-OV-7BBackbone=N/A, Training-free=-2025.12 | 46.4 | 56.5 | — | 56.5 | 15 | 51.5 | — | — | 52.5 | — | — | — | — | — | — | |
| Video-R12026.04 | 45.4 | — | — | — | — | — | — | — | — | 23.5 | — | 54.5 | 62 | 39.5 | 47.5 | |
| InternVL3.5-14BBase Model=InternVL3.5-14B2026.04 | 45.12 | — | 67.61 | — | — | — | — | — | — | 9.5 | 43.5 | 50 | 55 | — | — | |
| DINO-HEALBase Model=InternVL3.5-14B2026.04 | 45.11 | — | 67.05 | — | — | — | — | — | — | 10 | 43.5 | 50 | 55 | — | — | |
| VCDBase Model=InternVL3.5-14B2026.04 | 44.9 | — | 62.5 | — | — | — | — | — | — | 12 | 44.5 | 49.5 | 56 | — | — | |
| SELF-PEPBase Model=Video-LLaVA-7B2026.05 | 43.2 | — | — | — | — | — | — | — | — | 11 | 34 | 52 | 36 | 6.3 | — | |
| MotionCDBackbone=Qwen3-VL-4B-Instruct2026.06 | 42.79 | — | — | — | — | — | -0.19 | 0.24 | — | — | — | — | — | — | — | |
| SSCDBackbone=LLaVA-NeXT2026.01 | 35.4 | 60.5 | 32.4 | 52 | 18.5 | 44 | -0.04 | 0.44 | — | — | — | — | — | — | — | |
| LLaVA-NeXTBackbone=LLaVA-NeXT2026.01 | 32.3 | 58 | 21.6 | 47 | 15.5 | 46 | 0.16 | 0.73 | — | — | — | — | — | — | — | |
| TCDBackbone=LLaVA-NeXT2026.01 | 32.3 | 58.5 | 23.3 | 47 | 12.5 | 47 | 0.16 | 0.72 | — | — | — | — | — | — | — | |
| MotionCDBackbone=LLaVA-NeXT2026.01 | 24.8 | 48.5 | 20.5 | 34.5 | 13.5 | 28 | 0.06 | 0.58 | — | — | — | — | — | — | — | |
| MultiToPBackbone=Video-LLaVA-7B2026.06 | 24.08 | — | — | — | — | — | 0.29 | 0.86 | — | — | — | — | — | — | — | |
| MotionCDBackbone=Video-LLaVA-7B2026.06 | 23.16 | — | — | — | — | — | 0.27 | 0.81 | — | — | — | — | — | — | — | |
| SSCDBackbone=Video-LLaVA2026.01 | 22.9 | 46.5 | 29.5 | 21 | 8 | 30 | 0.2 | 0.75 | — | — | — | — | — | — | — | |
| MotionCDBackbone=Video-LLaVA2026.01 | 20.5 | 45 | 15.9 | 23.5 | 9.5 | 26 | 0.25 | 0.78 | — | — | — | — | — | — | — | |
| RegularBase Model=Video-LLaVA-7B2026.05 | 19.7 | — | — | — | — | — | — | — | — | 3 | 26 | 35.5 | 12 | 9.7 | — | |
| DeCoBase Model=Video-LLaVA-7B2026.05 | 19.7 | — | — | — | — | — | — | — | — | 3 | 26 | 35.5 | 12 | 9.7 | — | |
| DINO-HEALBackbone=Video-LLaVA-7B2026.06 | 18.14 | — | — | — | — | — | 0.37 | 0.94 | — | — | — | — | — | — | — | |
| TCDBackbone=Video-LLaVA-7B2026.06 | 17.21 | — | — | — | — | — | 0.38 | 0.94 | — | — | — | — | — | — | — | |
| Video-LLaVA-7BBackbone=Video-LLaVA-7B2026.06 | 16.39 | — | — | — | — | — | 0.38 | 0.94 | — | — | — | — | — | — | — | |
| Dino-HealBackbone=Video-LLaVA2026.01 | 15 | 36 | 10.2 | 12 | 4 | 25 | 0.36 | 0.92 | — | — | — | — | — | — | — | |
| SmartSightBackbone=Video-LLaVA-7B2026.06 | 15 | — | — | — | — | — | 0.35 | 0.97 | — | — | — | — | — | — | — | |
| TCDBackbone=Video-LLaVA2026.01 | 14.6 | 36 | 10.2 | 9.5 | 4 | 25.5 | 0.36 | 0.92 | — | — | — | — | — | — | — | |
| TA-AEBackbone=Video-LLaVA-7B2026.06 | 14.45 | — | — | — | — | — | 0.33 | 0.92 | — | — | — | — | — | — | — | |
| Video-LLaVABackbone=Video-LLaVA2026.01 | 14.3 | 35.5 | 9.7 | 11 | 3 | 24 | 0.36 | 0.92 | — | — | — | — | — | — | — | |
| Dino-HealBackbone=LLaVA-NeXT2026.01 | 12.4 | 6 | 22.7 | 2.5 | 18 | 25 | -0.11 | 0.38 | — | — | — | — | — | — | — |