MOEVE - Topic Extraction: leaderboard
MOEVE (Modelle fuer die Oeffentliche Verwaltung Evaluieren, Bundesdruckerei, 2026) benchmarks language models for the German public sector with German prompts and German datasets, many of them from public administration. This board is the mean overall topic-extraction score on a 0-100 scale over two German datasets, combining semantic topic matching with LLM-judged topic adherence; higher is better.
Metric: Mean overall topic-extraction score x 100 (0-100) over two German datasets (KIKC Topics, German Ministry Publications), combining semantic topic match with LLM-judged topic adherence; German prompts, default decoding, open-weight models on vLLM or Ollama at the quantization listed in the paper's model table, others by API; the paper prints only the top 10 of 39 evaluated models; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 64.1 | #741 |
| 2 | Mistral Small 3.1 | 64 | #1689 |
| 3 | GPT-4o Mini | 63.8 | #1623 |
| 4 | Llama 3.3 70B | 63.8 | #1498 |
| 5 | Mistral Large 3 | 62.5 | #817 |
| 6 | DeepSeek R1 Distill Qwen 32B | 62.5 | #1788 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=moeve-topic-extraction · How It Works · Data refreshed daily, snapshot 2026-11-13.