RoboProcessBench - Motion State: leaderboard

Metric: Accuracy (%) on 896 questions asking whether the manipulator is actively moving or effectively stationary (2 options, chance 50.0); zero-shot multiple-choice VQA on held-out robot manipulation recordings from GM-100, RH20T, REASSEMBLE and AIST-Bimanual (ProcessData-Eval), temperature 0.01; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 14 models tracked.

Top models

#ModelScore
1Qwen 3 VL 32B69.2
2GPT-4o68.4
3Claude Sonnet 4.668.1
4GPT-5.4 Mini67
5Claude Haiku 4.561.3
6InternVL3-38B58.1
7Qwen 2.5 VL 7B Instruct58
8InternVL3-8B54
9InternVL3.5-8B53.8
10GLM-4.6V48.9

Interactive version: theaggregate.ai/benchmark?slug=roboprocessbench-motion-state · How It Works · Data refreshed daily, snapshot 2026-09-29.