Volare: benchmark results
Moxoff's Italian-language SFT/LoRA fine-tune of Google's Gemma-7B, trained on SQuAD-it and in-house data for RAG and context-heavy tasks. Provider: Moxoff. Released 2024-04-15. Access: Open.
Aggregate Rating 1403 ± 22, rank #1213 of 1656 rated models, from 25 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - MMLU | 65.06 | Accuracy (%) (5-shot) | 81.1 |
| Pinocchio Italian - Lingua Straniera | 72.56 | Accuracy (%) | 79.5 |
| Pinocchio Italian - Matematica E Scienze | 50.96 | Accuracy (%) | 75 |
| Open LLM Leaderboard v1 - GSM8K | 52.31 | Accuracy (%) (5-shot) | 64.2 |
| Pinocchio Italian Leaderboard | 55.14 | Average Accuracy (%) | 61.4 |
| Open LLM Leaderboard v1 - WinoGrande | 78.14 | Accuracy (%) (5-shot) | 60.9 |
| Open LLM Leaderboard v1 - ARC Challenge | 62.97 | Normalized accuracy (%) (25-shot) | 60.6 |
| Pinocchio Italian - Cultura | 58.69 | Accuracy (%) | 59.1 |
| Pinocchio Italian - Generale | 55.83 | Accuracy (%) | 59.1 |
| Pinocchio Italian - Logica | 41.83 | Accuracy (%) | 59.1 |
| Pinocchio Italian - Diritto | 50.97 | Accuracy (%) | 54.5 |
| Open Italian LLM - MMLU-Pro (IT) | 23.61 | Accuracy (%) | 53.6 |
Interactive version: theaggregate.ai/model?slug=volare · How It Works · Data refreshed daily, snapshot 2026-10-20.