Skip to content

Benchmark

Arena ELO Image Edit

People compare model answers and choose which they prefer. Higher scores are better.

Models measured
58
Scores recorded
64
Sources
1
IQ weight
1.04

Score distribution

1,297
Mean
1,302.1
Median
113.8
Std dev
1,150.2
10th percentile
1,437.4
90th percentile
996.1
Min
996.11,518.6
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.04 for image-generation models, measured over 46 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1GPT-Image-2.5 Sunburst· LMArena data1,518.6
2GPT-Image-2.5 Flare· LMArena data1,483.6
3GPT-Image-2.6 SunburstMedium· LMArena data1,480.7
4MAI Image 2.7· LMArena data1,475.5
5GPT Image 2Medium· LMArena data1,461.5
6GPT-Image-2.6 FlareMedium· LMArena data1,460.7
7Grok Imagine Image 2.0· LMArena data1,437.4
8MAI-Image-2.6· LMArena data1,427.4
9Gemini Nano Banana 2.1· LMArena data1,427.2
10Grok Imagine Image 2.020260801· LMArena data1,418.6
11Krea 3 Ultra· LMArena data1,411.7
12Mai Image 2.5· LMArena data1,400.3
13Muse Image· LMArena data1,400
14Seedream 5.0 Pro· LMArena data1,391.9
15Grok Imagine ImageQuality, 20260519· LMArena data1,391.2
16Gemini 3 Pro Imagenano-banana-pro, 2K· LMArena data1,390.8
17ChatGPT Image LatestHigh Fidelity, 20251216· LMArena data1,387.5
18Gemini 3.1 Flash Imagenano-banana-2, Web Search· LMArena data1,386.9
19Gemini 3 Pro Imagenano-banana-pro· LMArena data1,384.8
20Reve 2.1· LMArena data1,372.8
21GPT Image 1.5High Fidelity· LMArena data1,367.3
22Qwen Image 2.1· LMArena data1,363.2
23Reve 2.0· LMArena data1,359.1
24MiniMax H3-Image· LMArena data1,358.2
25Ideogram V4.5· LMArena data1,346.8
26UNI-1.1 Max· LMArena data1,332.9
27Grok Imagine Image· LMArena data1,328.6
28UNI-1.1· LMArena data1,315.3
29Gemini 3.1 Flash-Lite Imagenano-banana-2-lite· LMArena data1,312.7
30Qwen Image 2.0 Pro2026-06-22· LMArena data1,302.8
31Doubao Seedream 4.5· LMArena data1,302.5
32Wan 2.7 Image Pro· LMArena data1,302.1
33HunyuanImage 3.0 Instruct· LMArena data1,301.4
34Wan 2.7 Image· LMArena data1,300.8
35Seedream 5.0 Lite· LMArena data1,293.4
36Gemini 2.5 Flash Imagenano-banana· LMArena data1,291.6
37Doubao Seedream 4-02K· LMArena data1,267
38FLUX.2 [max]· LMArena data1,261.7
39Reve v1.1· LMArena data1,261.5
40Kling Image O1· LMArena data1,250.4
41FLUX.2 [pro]· LMArena data1,247.5
42Qwen Image Edit· LMArena data1,240.9
43Reve v1· LMArena data1,234.6
44Qwen Image Edit2511· LMArena data1,233.1
45Wan 2.6 Image· LMArena data1,231.2
46FLUX.2 [flex]· LMArena data1,224.4
47FLUX.2 [klein] 9B· LMArena data1,224.4
48FLUX.2 [dev]· LMArena data1,223.9
49Doubao Seedream 4-0High Res, fal· LMArena data1,215.7
50P Image Edit· LMArena data1,209.7

Showing 1–50 of 64 results.

Data credits: LMArena, CC BY 4.0 (licence)