Autonomous Driving Robustness Evaluation on CODA-LM 94 (test)
57.5General Perception ScoreGPT-4V
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| GPT-4V2026.04 | 57.5 | 56.3 | 63.3 | 59 | |
| CPO++2026.04 | 55.6 | 85.2 | 65.5 | 68.8 | |
| CODA-VLMVenue=WACV’252026.04 | 55 | 77.7 | 58.1 | 63.6 | |
| Percept-DriveLMVenue=ICRA’252026.04 | 54.3 | 83.1 | 60.5 | 66 | |
| CPOVenue=NeurIPS’252026.04 | 48.2 | 82.3 | 60.9 | 63.5 | |
| MPDriveVenue=CVPR’252026.04 | 41.8 | 76.1 | 58.2 | 58.7 | |
| InternVLVenue=CVPR’242026.04 | 38.4 | 61.5 | 41.2 | 47 | |
| Qwen-VL-MAXVenue=Commercial2026.04 | 34.6 | 68.2 | 47.4 | 50.1 | |
| LLaVA-NeXTVenue=ICLR’252026.04 | 29.9 | 53.6 | 31.9 | 38.5 | |
| Gemini-ProVenue=Commercial2026.04 | 25.2 | 51.4 | 27.4 | 34.7 |