Benchmark
Arena ELO Text to Image - Chinese
Higher scores are better.
All image generation benchmarksScores from LMArena data.
- Models measured
- 8
- Scores recorded
- 8
- Sources
- 1
- IQ weight
- 0.54
Score distribution
- 868.7
- Mean
- 978
- Median
- 137
- Std dev
- 719.9
- 10th percentile
- 1,024.6
- 90th percentile
- 719.2
- Min
This benchmark has a weight of 0.54 for image-generation models, measured over 8 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 |
|---|---|---|
| 1 | Recraft V3· LMArena data | 1,029 |
| 2 | Luma Photon· LMArena data | 1,024.6 |
| 3 | Ideogram V2· LMArena data | 983.1 |
| 4 | DALL-E 3· LMArena data | 978 |
| 5 | Imagen 3· LMArena data | 775.4 |
| 6 | FLUX.1 [dev]· LMArena data | 720.7 |
| 7 | FLUX1.1 [pro]· LMArena data | 719.9 |
| 8 | Stable Diffusion 3.5 Large· LMArena data | 719.2 |
Data credits: LMArena, CC BY 4.0 (licence)