AA Global-MMLU-Lite - German: leaderboard
German-language slice of Artificial Analysis Global-MMLU-Lite, evaluating MMLU-style knowledge and reasoning across STEM, humanities, and social-science subjects.
Metric: Accuracy (%). Source: artificialanalysis.ai. Saturation forecast: Estimated already saturated. 111 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 94.75 | #58 |
| 2 | Claude Opus 4.6 (Max) | 93.42 | #63 (Claude Opus 4.6) |
| 3 | Gemini 3 Pro (Preview) (High) | 93.17 | #66 (Gemini 3 Pro (Preview)) |
| 4 | Claude Opus 4.5 (Thinking) | 93 | #83 (Claude Opus 4.5) |
| 5 | GPT-5 (High) | 92.5 | #98 (GPT-5) |
| 6 | Gemini 2.5 Pro | 92.42 | #152 |
| 7 | GPT-5 (Medium) | 92.33 | #98 (GPT-5) |
| 8 | Claude Opus 4.5 (Non-reasoning) | 92.25 | #83 (Claude Opus 4.5) |
| 9 | Claude Sonnet 4.6 (Max) | 92.25 | #90 (Claude Sonnet 4.6) |
| 10 | GPT-5.1 (High) | 91.83 | #135 (GPT-5.1) |
| 11 | Grok 4.20 0309 (Reasoning) | 91.75 | #112 (Grok 4.20 0309) |
| 12 | Claude Sonnet 4.5 (Thinking) | 91.67 | #142 (Claude Sonnet 4.5) |
| 13 | Qwen 3.5 397B A17B (Reasoning) | 91.67 | #146 (Qwen 3.5 397B A17B) |
| 14 | Qwen 3.5 122B A10B (Reasoning) | 91.58 | #173 (Qwen 3.5 122B A10B) |
| 15 | Grok 4 | 91.5 | #181 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=aa-global-mmlu-lite-german · How It Works · Data refreshed daily, snapshot 2026-10-20.