CrashSight - Involved Parties: leaderboard

Involved-party questions (vehicle and vulnerable road user types, features and counts) of CrashSight (2026), a multiple-choice VQA benchmark on real roadside-camera traffic crash videos answered zero-shot from four frames. Accuracy in percent; higher is better.

Metric: Accuracy (%) on involved-party questions (vehicle and vulnerable-road-user types, features, counts), four-option multiple-choice VQA on held-out roadside-camera crash videos of CrashSight-VQA (4 frames at 1 fps, zero-shot, deterministic decoding, exact letter match); higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 6 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1InternVL3-8B58.1#1795
2LLaVA-NeXT-Video-7B52.3#5783
3InternVL3-2B52.1#3318
4Qwen 2.5 VL 7B Instruct51.9#1919
5Qwen 2.5 VL 3B Instruct50.6#2720
6LLaVA-OneVision-0.5B36#7977

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=crashsight-involved-parties · How It Works · Data refreshed daily, snapshot 2026-11-22.