EuroEval German NLU - Germanquad: leaderboard
The "Germanquad" subtask of EuroEval German NLU. German-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: Not forecast. 321 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma 2 27B | 69.98 | #806 |
| 2 | Llama 3.1 405B | 69.43 | #466 |
| 3 | Llama 2 70B Base | 69.16 | #1025 |
| 4 | Qwen3.5-4B-Base | 68.99 | #798 |
| 5 | gemma-3-12B-pt | 68.42 | #686 |
| 6 | Gemma 3n E4B | 68.37 | #878 |
| 7 | Yi 1.5 34B | 68.22 | #807 |
| 8 | Llama 3.1 70B | 68.12 | #592 |
| 9 | gemma-3-27B-pt | 67.98 | #502 |
| 10 | Bielik-11B-v2.3-Instruct | 67.05 | #693 |
| 11 | Yi-1.5-9B | 66.85 | #1241 |
| 12 | OLMo-2-1124-7B | 66.81 | #1401 |
| 13 | Qwen3.5-2B-Base | 65.71 | #1309 |
| 14 | Magistral Small 1.2 | 65.45 | #570 |
| 15 | Ministral-3-14B-Reasoning-2512 | 65.4 | #540 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-german-nlu-germanquad · How It Works · Data refreshed daily, snapshot 2026-10-20.