DriveSpatial - Reasoning: leaderboard

DriveSpatial (2026) tests whether vision-language models can build and reason over a coherent driving scene from synchronized multi-view camera frames, with human-verified questions generated from scene graphs of nuScenes, Waymo, TruckScenes, AV2 and ONCE. This board covers the spatiotemporal reasoning tasks on the full 15,670-question set, scored as accuracy in percent; higher is better.

Metric: Accuracy (%) on the spatiotemporal reasoning tasks (interactions and future dynamics) on the full benchmark (15,670 human-verified questions over 20 tasks from nuScenes, Waymo, TruckScenes, AV2 and ONCE); open-source generalist and driving or spatial specialist VLMs; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 12 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Qwen 3 VL 8B47.66#917
2InternVL3.5-8B41.31#1604
3Gemma 3 12B (IT)40.8#1887
4LLaVA-OneVision-7B40.34#2981
5RoboTron-Drive39.56
6DeepSeek-VL2-Small39.26#2464
7LLaVA-Video-7B36.89#2276
8LongVA-7B35.25#3456
9DriveSpatial Ego3D (checkpoint unspecified)34.53
10DriveSpatial SenseNova-SI (checkpoint unspecified)32.73
11SpaceThinker31.65#1448
12InternVL3-8B30.65#1782

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=drivespatial-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-21.