Benchmark
Arena ELO Text Factuality - German
People compare model answers and choose which they prefer. Higher scores are better.
All human preference benchmarksScores from LMArena data.
- Models measured
- 49
- Scores recorded
- 55
- Sources
- 1
- IQ weight
- 1.57
Score distribution
- 1,417.9
- Mean
- 1,417.8
- Median
- 29
- Std dev
- 1,383.5
- 10th percentile
- 1,462
- 90th percentile
- 1,363.6
- Min
This benchmark has a weight of 1.57 for chat models, measured over 49 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 51 | Inkling Small· LMArena data | 1,380.7 |
| 52 | Kimi K2 Thinking· LMArena data | 1,376 |
| 53 | Doubao Seed 2.0 Pro· LMArena data | 1,374 |
| 54 | Claude Haiku 4.520251001· LMArena data | 1,369.4 |
| 55 | Qwen3.5 Flash· LMArena data | 1,363.6 |
Showing 51–55 of 55 results.
Data credits: LMArena, CC BY 4.0 (licence)