EuroEval Dutch Summarization - Wiki Lingua NL: leaderboard
EuroEval's Dutch summarization task on the Wiki Lingua NL dataset — models write abstractive summaries of Dutch news articles, scored with automatic text-overlap metrics, from the EuroEval European language model benchmark suite.
Metric: chrF3++ (%). Source: euroeval.com. Saturation forecast: Not forecast. 151 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Apertus-70B-Instruct-2509 | 36.71 | #820 |
| 2 | Mistral Small 3.1 | 36.35 | #610 |
| 3 | Qwen 3 14B (Non-reasoning) | 36.05 | #537 (Qwen 3 14B) |
| 4 | Llama 3.1 8B Instruct | 36.04 | #1035 |
| 5 | Mistral Small 3.2 | 35.89 | #605 |
| 6 | Llama 3.3 70B Instruct | 35.72 | #534 |
| 7 | GPT-4.1 | 35.58 | #248 |
| 8 | Gemma 3n E4B (IT) | 35.48 | #749 |
| 9 | Gemma 3 4B (IT) | 35.47 | #992 |
| 10 | Qwen 3 30B A3B 2507 Instruct | 35.22 | #488 |
| 11 | Qwen 3 32B (Non-reasoning) | 35.22 | #443 (Qwen 3 32B) |
| 12 | Olmo-3.1-32B-Instruct-DPO | 35.07 | #845 |
| 13 | Olmo 3.1 32B Instruct | 34.91 | #763 |
| 14 | Qwen 2.5 72B Instruct | 34.9 | #454 |
| 15 | Gemma 3 12B (IT) | 34.88 | #664 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-dutch-summarization-wiki-lingua-nl · How It Works · Data refreshed daily, snapshot 2026-10-20.