DashBench CCW Subset - Q2: leaderboard

DashBench evaluates video understanding on 1,709 driving clips. Five explanation prompts and binary event detection are scored separately; mean agreement excludes the binary task. The 823-clip corner-case-window subset and training-free decoder variants retain their own conditions. Q5 asks for definition-guided corner-case identification and description. This board reports lLM-judged answer accuracy (%). Higher is better.

Metric: LLM-judged answer accuracy (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 3 models tracked. Updated 14 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-5Evaluation run: DashBench CCW823 CCW 16 frames42.41#167
2Qwen3-VL (8B)Evaluation run: DashBench CCW823 CCW 32 frames20.66
3InternVL3.5 (8B)Evaluation run: DashBench CCW823 CCW 32 frames12.88

Interactive version: theaggregate.ai/benchmark?slug=dashbench-ccw-subset-q2 · How It Works · Data refreshed daily, snapshot 2026-11-16.