RoboBrain-2.0-7B: benchmark results
Provider: Other.
Aggregate Rating 1535 ± 31, rank #1889 of 10916 rated models, from 61 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| ESPIRE - Place Localization | 50.7 | Localization accuracy (%) on the 1,125 place tasks: the pred | 100 |
| RoboProcessBench - Outcome Prediction | 62.5 | Accuracy (%) on 793 questions asking whether the ongoing att | 100 |
| RoboProcessBench - Primitive Chain Restoration | 91.3 | Accuracy (%) on 46 questions asking which primitive fills a | 96.2 |
| ESPIRE - Pick Localization | 57.72 | Localization accuracy (%) on the 1,095 pick tasks: the predi | 90 |
| RoboProcessBench - Motion Direction | 44.4 | Accuracy (%) on 1,207 questions asking the dominant motion d | 76.9 |
| NutriBench-Kitchen - Long-Term Planning (Hard) | 58 | Score (%; the long-term planning task family, hard regime: d | 72.7 |
| EPIC-Bench - Ground Detection | 42.92 | Navigation score (0-100) on ground detection (localizing tra | 70.5 |
| NutriBench-Kitchen - Short-Term Planning (Easy) | 53.3 | Score (%; the short-term planning task family, easy regime: | 68.2 |
| NutriBench-Kitchen - Long-Term Planning (Easy) | 86 | Score (%; the long-term planning task family, easy regime: s | 63.6 |
| SiT-Bench - Multi-View & Geometric Reasoning | 46.17 | Accuracy (%; item-weighted over the category's subtasks) | 61.7 |
| Causal-Plan-Bench - State Evolution | 38.4 | Open-ended state evolution (the ongoing action and its immed | 61.5 |
| SiT-Bench - Embodied & Fine-grained Perception | 41.36 | Accuracy (%; item-weighted over the category's subtasks) | 60 |
Interactive version: theaggregate.ai/model?slug=robobrain-2-0-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.