EuroEval Italian NLU - MultiNERD IT — leaderboard
Metric: Named entity recognition Score (%). Source: euroeval.com. 429 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 4B 2507 (Thinking) | 87.46 |
| 2 | Llama 3.1 405B Instruct FP8 | 85.35 |
| 3 | Qwen 3 Next 80B A3B (Thinking) | 85.32 |
| 4 | Qwen 3 30B A3B | 84.97 |
| 5 | Llama 3 70B | 84.06 |
| 6 | Llama 3.1 70B Instruct | 83.94 |
| 7 | Claude Sonnet 4.5 (Thinking) | 83.37 |
| 8 | Llama 3.3 70B Instruct | 83.25 |
| 9 | Llama 3.1 70B | 82.88 |
| 10 | QwQ-32B | 82.7 |
| 11 | Mistral Small 3.2 | 82.6 |
| 12 | Magistral Small | 81.32 |
| 13 | GPT-5.4 Mini (High) | 80.88 |
| 14 | Llama 3 70B Instruct | 80.71 |
| 15 | Gemma 4 31B (IT) | 80.41 |
Interactive version: theaggregate.ai/benchmark?slug=euroeval-italian-nlu-multinerd-it · How the rankings work · Data refreshed daily, snapshot 2026-07-22.