Virtuoso-Medium-v2: benchmark results
An Arcee AI Qwen2.5-32B derivative distilled from DeepSeek-V3 logits with tokenizer alignment, fusion merging and DPO. Provider: Arcee AI. Access: Open.
Aggregate Rating 1559 ± 33, rank #1561 of 10916 rated models, from 55 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Japanese LLM - Jmmlu Exact Match | 76.65 | Score (%) | 97.6 |
| Open Japanese LLM - HE | 78.63 | Score (%) | 97.1 |
| Open Japanese LLM - Wiki NER SET F1 | 15.93 | Score (%) | 96.9 |
| Open Japanese LLM - Wiki PAS SET F1 | 11.78 | Score (%) | 96.9 |
| Open Japanese LLM - Jcommonsenseqa Exact Match | 96.43 | Score (%) | 96 |
| Open Japanese LLM - NLI | 81.62 | Score (%) | 95.7 |
| Open Japanese LLM - MMLU EN Exact Match | 80.62 | Score (%) | 95 |
| Open Japanese LLM - Jamp Exact Match | 70.69 | Score (%) | 92.9 |
| Open Japanese LLM - Jsts Spearman | 87.98 | Score (%) | 92.9 |
| Open Japanese LLM - Kuci Exact Match | 83.68 | Score (%) | 92.2 |
| Open Japanese LLM - FA | 30.35 | Score (%) | 91.6 |
| Open Japanese LLM - MC | 90.28 | Score (%) | 91.6 |
Interactive version: theaggregate.ai/model?slug=virtuoso-medium-v2 · How It Works · Data refreshed daily, snapshot 2026-11-08.