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
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 78.4 |
| 2 | Gemini 3.1 Pro (Preview) | 62.7 |
| 3 | Claude Opus 4.7 | 56.6 |
| 4 | DeepSeek V4 Pro | 49.1 |
| 5 | GLM-5.2 | 43.6 |
| 6 | Qwen 3.7 Plus | 24.3 |
| 7 | GPT-5.4 | 23.1 |
| 8 | Qwen 3.6 Plus | 21.1 |
| 9 | Qwen 3 VL 8B | 1.5 |
Interactive version: theaggregate.ai/benchmark?slug=robographbench-tabletop · How It Works · Data refreshed daily, snapshot 2026-09-26.