EuroEval German NLU: leaderboard
German-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: NLU Average Score (%). Source: euroeval.com. Saturation forecast: Estimated already saturated. 327 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.1 70B | 65.97 | #592 |
| 2 | GPT-5.4 Mini (2026-03-17) (High) | 63.89 | #169 (GPT-5.4 Mini (2026-03-17)) |
| 3 | Gemma 4 31B (IT) | 63.85 | #176 |
| 4 | Gemini 3 Pro (Preview) | 63.61 | #66 |
| 5 | Mistral Small 3.2 | 62.48 | #605 |
| 6 | Claude Sonnet 4.6 | 62.17 | #90 |
| 7 | Mistral Small 3.1 | 62.02 | #610 |
| 8 | Gemma 4 26B A4B (IT) | 61.81 | #297 |
| 9 | Gemini 3.1 Flash Lite (Preview) | 61.63 | #194 |
| 10 | Llama 3.1 405B | 61.22 | #466 |
| 11 | gemma-3-27B-pt | 61.05 | #502 |
| 12 | Qwen 3 235B A22B 2507 FP8 (Thinking) | 61.02 | #246 (Qwen 3 235B A22B 2507 FP8) |
| 13 | Magistral Small 1.2 | 60.72 | #570 |
| 14 | Llama 3.1 405B Instruct FP8 | 60.37 | #389 |
| 15 | Ministral-3-14B-Reasoning-2512 | 60.21 | #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 · How It Works · Data refreshed daily, snapshot 2026-10-20.