RoboBrain2.5-8B-NV: benchmark results
Provider: Other. Access: Open.
Aggregate Rating 1543 ± 31, rank #1794 of 10916 rated models, from 44 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Causal-Plan-Bench - Spatial Precondition | 42 | Accuracy (%) on the spatial precondition (conditions that mu | 96.2 |
| Causal-Plan-Bench - Affordance Visual Semantics | 35 | Accuracy (%) on the affordance visual semantics (manipulated | 92.3 |
| Causal-Plan-Bench - Counterfactual Outcome | 38.9 | Open-ended counterfactual outcome (the consequence if a key | 92.3 |
| Causal-Plan-Bench - Executability | 44 | Executability score: mean four-choice accuracy (%) after ans | 92.3 |
| Causal-Plan-Bench - State Evolution | 41.6 | Open-ended state evolution (the ongoing action and its immed | 92.3 |
| Causal-Plan-Bench - Affordance Precondition | 46 | Accuracy (%) on the affordance precondition (functional obje | 84.6 |
| Causal-Plan-Bench - Composition | 34 | Composition score: mean rubric score (0-100): a blind GPT-5. | 84.6 |
| Causal-Plan-Bench - Physical Feasibility | 44 | Accuracy (%) on the physical feasibility (why the action is | 84.6 |
| OVO-S-Bench - Generative Spatial Reasoning | 50.63 | Level 3 (generative spatial reasoning: simulation, consisten | 70 |
| SpatialGen-Bench (Text Answering) | 47.31 | Overall Accuracy (%) | 65.4 |
| Causal-Plan-Bench | 36.57 | Overall score (0-100): macro-average of the 12 task scores, | 61.5 |
| Causal-Plan-Bench - Affordance Postcondition | 37 | Accuracy (%) on the affordance postcondition (functional sta | 61.5 |
Interactive version: theaggregate.ai/model?slug=robobrain2-5-8b-nv · How It Works · Data refreshed daily, snapshot 2026-11-08.