Drive-P2D - Optimal Action (Decision-1): leaderboard
Metric: Accuracy (%; select the ego vehicle's optimal action (single choice); zero-shot, all scenarios, mean of two runs). Source: arxiv.org. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4.1 | 56.95 |
| 2 | Qwen 2.5 VL 72B Instruct | 54.69 |
| 3 | Gemma 3 27B (IT) | 50.4 |
| 4 | Gemini 2.5 Pro | 49.47 |
| 5 | Gemma 3 4B (IT) | 45.78 |
| 6 | Llama 3.2 90B Vision Instruct | 45.14 |
| 7 | Phi-4 Multimodal Instruct | 44.58 |
| 8 | Llama 3.2 11B Instruct | 41.84 |
| 9 | Qwen 2.5 VL 7B Instruct | 40.87 |
Interactive version: theaggregate.ai/benchmark?slug=drive-p2d-optimal-action-decision-1 · How It Works · Data refreshed daily, snapshot 2026-09-26.