EuroEval Italian NLU - SQuAD IT: leaderboard
The "SQuAD IT" subtask of EuroEval Italian NLU. Italian-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: Character F1 (%). Source: euroeval.com. Saturation forecast: Around January 2027. 429 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Ministral-3-14B-Reasoning-2512 | 78.42 | #540 |
| 2 | Qwen3.5-4B-Base | 78.05 | #798 |
| 3 | SOLAR-10.7B-v1.0 | 77.48 | #1253 |
| 4 | Magistral Small 1.2 | 76.71 | #570 |
| 5 | Mistral Small 3.2 | 76.1 | #605 |
| 6 | Llama 3.1 405B | 75.85 | #466 |
| 7 | Ministral-3-14B-Instruct-2512 | 75.14 | #615 |
| 8 | Claude Sonnet 4.5 (Non-reasoning) | 75.1 | #142 (Claude Sonnet 4.5) |
| 9 | Gemini 2.5 Flash Lite (Non-reasoning) | 74.89 | #422 (Gemini 2.5 Flash Lite) |
| 10 | tiny-aya-water | 74.89 | |
| 11 | Gemma 2 27B | 74.8 | #806 |
| 12 | Mistral Small 3.1 | 74.76 | #610 |
| 13 | Yi 1.5 34B | 74.69 | #807 |
| 14 | Llama 3.1 70B | 74.66 | #592 |
| 15 | Tiny Aya Global | 74.62 | #1325 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-italian-nlu-squad-it · How It Works · Data refreshed daily, snapshot 2026-10-20.