RoboGraphBench - Tabletop: leaderboard

Metric: Task success rate (%; episodes that reach the goal and stop correctly, averaged over the baseline and six intervention conditions; semantic closed-loop evaluation on partially observed symbolic scene graphs, at most 100 steps; 399 tabletop episodes from 57 scenes). Source: arxiv.org. Saturation forecast: Around December 2026. 15 models tracked.

Top models

#ModelScore
1GPT-5.578.4
2Gemini 3.1 Pro (Preview)62.7
3Claude Opus 4.756.6
4DeepSeek V4 Pro49.1
5GLM-5.243.6
6Qwen 3.7 Plus24.3
7GPT-5.423.1
8Qwen 3.6 Plus21.1
9Qwen 3 VL 8B1.5

Interactive version: theaggregate.ai/benchmark?slug=robographbench-tabletop · How It Works · Data refreshed daily, snapshot 2026-09-26.