Belebele European - Dutch: leaderboard
Metric: Accuracy (%, 0-shot). Source: huggingface.co. 29 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2.5 14B Instruct | 89.89 |
| 2 | Gemma 3 12B (IT) | 88.33 |
| 3 | Qwen 3 14B | 88.22 |
| 4 | Gemma 2 9B (IT) | 87.33 |
| 5 | Qwen 3 8B | 86.44 |
| 6 | GaMS3-12B-Instruct | 85.67 |
| 7 | Phi-4 | 85.33 |
| 8 | Qwen 2.5 7B | 80.89 |
| 9 | Mistral Nemo Instruct (2407) | 79.22 |
| 10 | Gemma 2 9B | 69.78 |
| 11 | Apertus-8B-Instruct-2509 | 69.56 |
| 12 | Mistral 7B Instruct (v0.2) | 56.11 |
| 13 | PLLuM-12B-nc-chat-250715 | 54.56 |
| 14 | Mistral-7B-v0.2 | 50.56 |
| 15 | Bielik-11B-v2 | 49.78 |
Interactive version: theaggregate.ai/benchmark?slug=belebele-european-dutch · How It Works · Data refreshed daily, snapshot 2026-09-19.