NEO-9B-SFT: benchmark results
Provider: Other.
Aggregate Rating 1471 ± 42, rank #2820 of 10916 rated models, from 19 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| VISTA-Bench (Pure Text) | 59.5 | Accuracy (%) over the same 1,500 questions given as ordinary | 51.7 |
| PLAN-Bench - Spatial Cognition | 23.6 | Accuracy (%) | 48.1 |
| Game2Real-Bench - Toy Game - Fine-grained Perception | 13.5 | Accuracy (%) | 37.5 |
| Game2Real-Bench - Real World - Fine-grained Perception | 32 | Task score (%) | 35.4 |
| PLAN-Bench - Overall | 18.6 | Accuracy (%) | 33.3 |
| Game2Real-Bench - Toy Game - Sequential Planning | 16.6 | Accuracy (%) | 29.2 |
| Game2Real-Bench - Real World - Constraint Reasoning | 21.6 | Task score (%) | 20.8 |
| Game2Real-Bench - Real World - Sequential Planning | 30 | Task score (%) | 20.8 |
| PLAN-Bench - Temporal Reasoning | 19.1 | Accuracy (%) | 20.4 |
| Game2Real-Bench - Toy Game - Spatial Reasoning | 12.6 | Accuracy (%) | 16.7 |
| PLAN-Bench - Multi-Step Planning | 12.6 | Accuracy (%) | 16.7 |
| Game2Real-Bench - Overall | 17 | Mean per-sample score (%) | 12.5 |
Interactive version: theaggregate.ai/model?slug=neo-9b-sft · How It Works · Data refreshed daily, snapshot 2026-11-08.