Pairwise progress-judgment on RoboPulse Small hop range
66Accuracy (Real)Robo-Dopamine
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Robo-Dopaminezero-shot=true2026.03 | 66 | 89 | 87 | 76 | 80 | |
| GVLzero-shot=true2026.03 | 61 | 67 | 58 | 67 | 63 | |
| VLACzero-shot=true2026.03 | 60 | 61 | 66 | 57 | 61 | |
| CLIP ViT-B/32 (I2I)zero-shot=true, backbone=ViT-B/32, input_interface=I2I2026.03 | 56 | 52 | 58 | 58 | 56 | |
| Gemini 3 Pro Previewzero-shot=true2026.03 | 55 | 62 | 43 | 56 | 54 | |
| CLIP ViT-L/14 (I2I)zero-shot=true, backbone=ViT-L/14, input_interface=I2I2026.03 | 52 | 56 | 55 | 56 | 54 | |
| CLIP ViT-L/14 (T2I)zero-shot=true, backbone=ViT-L/14, input_interface=T2I2026.03 | 50 | 46 | 48 | 52 | 49 | |
| Qwen3-VL-8B-Instructzero-shot=true, parameters=8B, variant=Instruct2026.03 | 49 | 51 | 44 | 47 | 48 | |
| CLIP ViT-B/32 (T2I)zero-shot=true, backbone=ViT-B/32, input_interface=T2I2026.03 | 48 | 49 | 56 | 49 | 51 | |
| Qwen3-VL-4B-Instructzero-shot=true, parameters=4B, variant=Instruct2026.03 | 47 | 49 | 34 | 53 | 46 | |
| GPT-5.2zero-shot=true2026.03 | 46 | 46 | 47 | 49 | 47 |