Urgency Prediction on VecEval (val)
12.12Intervention Failure RateFleetAgent
Evaluation Results
| Method | Links | |
|---|---|---|
| FleetAgentInput Modality=Vectorized V2N Messages2026.06 | 12.12 | |
| FleetAgent (w/o tokens prioritization)Input Modality=Vectorized V2N Messages2026.06 | 12.42 | |
| GPT-4oInput Modality=Language Description2026.06 | 13.63 | |
| Qwen-2.5VL-7BInput Modality=BEV Images, Evaluation Protocol=Few-shot Example2026.06 | 13.83 | |
| Qwen-2.5VL-7BInput Modality=Language Description, Evaluation Protocol=Few-shot Example2026.06 | 15.14 | |
| GPT-4oInput Modality=Raw Images2026.06 | 15.24 | |
| Qwen-2.5VL-7BInput Modality=Raw Images, Evaluation Protocol=Few-shot Example2026.06 | 15.9 | |
| Gemini-2.5-Flash (Reference Baseline)Input Modality=Language Description2026.06 | 16 | |
| GPT-4oInput Modality=BEV Images2026.06 | 16.05 |