Virtuoso-Small: benchmark results
Arcee AI Virtuoso-Small is a Qwen2.5-14B-based instruction model, distinct from the later DeepSeek-distilled v2 release. Provider: Arcee AI. Access: Open.
Aggregate Rating 1581 ± 26, rank #1236 of 10916 rated models, from 69 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard - IFEval | 79.35 | Score | 96.6 |
| Open Portuguese LLM - ENEM | 82.09 | Accuracy (%) | 95.5 |
| Open Portuguese LLM - OAB Exams | 62.87 | Accuracy (%) | 95.5 |
| Open Portuguese LLM - BLUEX | 74.55 | Accuracy (%) | 95.3 |
| Open Portuguese LLM - ASSIN2 RTE | 94.11 | Macro F1 (%) | 94.8 |
| Open LLM Leaderboard - BBH | 65.18 | Score | 91.7 |
| Open LLM Leaderboard - MATH Level 5 | 40.94 | Score | 90.8 |
| Open LLM Leaderboard - MMLU-Pro | 51.91 | Score | 90 |
| Open Japanese LLM - Wikicorpus E TO J Comet Wmt22 | 84.21 | Score (%) | 84.9 |
| Open Japanese LLM - Wiki Coreference SET F1 | 7.53 | Score (%) | 83.7 |
| Open LLM Leaderboard - GPQA | 33.64 | Score | 83.1 |
| Open Japanese LLM - Jcommonsenseqa Exact Match | 95.71 | Score (%) | 81.6 |
Interactive version: theaggregate.ai/model?slug=virtuoso-small · How It Works · Data refreshed daily, snapshot 2026-11-08.