EuroEval Dutch NLU - CoNLL NL: leaderboard
The "CoNLL NL" subtask of EuroEval Dutch NLU. Dutch-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: Micro F1 excluding MISC (%). Source: euroeval.com. Saturation forecast: Around January 2027. 332 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | OLMo 3.1 32B (Thinking) | 81.64 | #632 (OLMo 3.1 32B) |
| 2 | NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 | 80.79 | #679 |
| 3 | Gemma 4 31B (IT) | 80.2 | #176 |
| 4 | Qwen 3 4B 2507 (Thinking) | 80.14 | #538 (Qwen 3 4B 2507) |
| 5 | Qwen 3 235B A22B 2507 FP8 (Thinking) | 79.8 | #246 (Qwen 3 235B A22B 2507 FP8) |
| 6 | Qwen 3.5 35B A3B FP8 | 79.32 | #362 |
| 7 | Claude Sonnet 4.5 (Thinking) | 77.39 | #142 (Claude Sonnet 4.5) |
| 8 | Gemini 3 Flash (Preview) (Thinking) | 75.26 | #82 (Gemini 3 Flash (Preview)) |
| 9 | Qwen 3 Next 80B A3B (Thinking) | 75.09 | #320 (Qwen 3 Next 80B A3B) |
| 10 | QwQ-32B | 75.09 | #421 |
| 11 | Voxtral-Small-24B-2507 | 74.99 | #581 |
| 12 | Claude Sonnet 4.5 (Non-reasoning) | 74.97 | #142 (Claude Sonnet 4.5) |
| 13 | Llama 3.1 405B | 74.64 | #466 |
| 14 | Qwen 3.5 35B A3B Base | 74.6 | #310 |
| 15 | Gemini 3 Flash (Preview) (Non-reasoning) | 74.4 | #82 (Gemini 3 Flash (Preview)) |
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-nlu-conll-nl · How It Works · Data refreshed daily, snapshot 2026-10-20.