EuroEval German: leaderboard
German-language EuroEval all-task suite spanning NLU, knowledge, common-sense reasoning, reading comprehension, summarization, and related tasks.
Metric: Average Score (%). Source: euroeval.com. Saturation forecast: Estimated already saturated. 330 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 235B A22B 2507 FP8 (Thinking) | 68.41 | #246 (Qwen 3 235B A22B 2507 FP8) |
| 2 | GPT-5 (2025-08-07) (High) | 67.11 | |
| 3 | Qwen 3 235B A22B 2507 Instruct | 66.96 | #300 |
| 4 | Llama 3.1 405B Instruct FP8 | 66.56 | #389 |
| 5 | O3 (2025-04-16) | 65.78 | #121 |
| 6 | GPT-5 | 65.56 | #98 |
| 7 | Gemini 3 Pro (Preview) | 65.52 | #66 |
| 8 | Claude 3.7 Sonnet (20250219) (Thinking) | 65.08 | #201 (Claude 3.7 Sonnet (20250219)) |
| 9 | Qwen 3 Next 80B A3B Instruct | 64.97 | #368 |
| 10 | Gemini 2.5 Pro | 64.72 | #152 |
| 11 | Llama 3.1 405B | 64.5 | #466 |
| 12 | GPT-5.4 Mini (2026-03-17) (High) | 64.21 | #169 (GPT-5.4 Mini (2026-03-17)) |
| 13 | Qwen 3 235B A22B | 63.74 | #313 |
| 14 | Gemma 4 31B (IT) | 63.57 | #176 |
| 15 | Claude Sonnet 4.6 | 63.43 | #90 |
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 · How It Works · Data refreshed daily, snapshot 2026-10-20.