Skip to content

Benchmark

Arena ELO Text to Image

Higher scores are better.

Models measured
79
Scores recorded
86
Sources
1
IQ weight
1.43

Score distribution

1,149.3
Mean
1,141.2
Median
113.5
Std dev
1,015.6
10th percentile
1,318.1
90th percentile
897.3
Min
897.31,428.5
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.43 for image-generation models, measured over 68 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 data1,428.5
2GPT-Image-2.6 SunburstMedium· LMArena data1,410.1
3GPT-Image-2.5 Flare· LMArena data1,399.5
4MAI Image 2.7· LMArena data1,387.4
5GPT Image 2Medium· LMArena data1,382
6GPT-Image-2.6 FlareMedium· LMArena data1,375.3
7MAI-Image-2.6· LMArena data1,330.7
8Gemini Nano Banana 2.1· LMArena data1,327.6
9Grok Imagine Image 2.0· LMArena data1,318.1
10Reve 2.1· LMArena data1,301.8
11Grok Imagine Image 2.020260801· LMArena data1,287
12Reve 2.0· LMArena data1,268.3
13Muse Image· LMArena data1,265.6
14Gemini 3.1 Flash Imagenano-banana-2, Web Search· LMArena data1,256.1
15Qwen Image 3.0 Pro· LMArena data1,254.3
16Mai Image 2.5· LMArena data1,252.6
17Seedream 5.0 Pro· LMArena data1,249.3
18Gemini 3.1 Flash-Lite Imagenano-banana-2-lite· LMArena data1,249.1
19Gemini 3 Pro Imagenano-banana-pro, 2K· LMArena data1,245.7
20GPT Image 1.5High Fidelity· LMArena data1,236.1
21Gemini 3 Pro Imagenano-banana-pro· LMArena data1,231.2
22Qwen Image 2.1· LMArena data1,227.3
23Ideogram V4Quality· LMArena data1,198.9
24MiniMax H3-Image· LMArena data1,198.2
25Qwen Image 2.0 Pro2026-06-22· LMArena data1,190.2
26UNI-1.1 Max· LMArena data1,188.4
27MAI-Image-2· LMArena data1,182.7
28UNI-1.1· LMArena data1,181.4
29Recraft V4.1 Utility Pro· LMArena data1,169.3
30Grok Imagine Image· LMArena data1,168.1
31Cosmos3-Super-Text2ImageAgentic· LMArena data1,164.1
32FLUX.2 [max]· LMArena data1,161
33Grok Imagine Image Pro· LMArena data1,160
34FLUX.2 [flex]· LMArena data1,156.4
35Reve v1.5· LMArena data1,154.1
36FLUX.2 [pro]· LMArena data1,152.5
37Hunyuan Image V3· LMArena data1,151.5
38Cosmos3-Super-Text2Image· LMArena data1,149.6
39Gemini 2.5 Flash Imagenano-banana· LMArena data1,148.3
40Doubao Seedream 4.5· LMArena data1,147.8
41Imagen 4 Ultragenerate-001· LMArena data1,145.9
42FLUX.2 [dev]· LMArena data1,145
43Seedream 5.0 Lite· LMArena data1,141.2
44Wan 2.6· LMArena data1,138.5
45Doubao Seedream 4-02K· LMArena data1,138.3
46Recraft V4.1 Pro· LMArena data1,132
47Krea v2 Medium· LMArena data1,129.7
48Imagen 4generate-001· LMArena data1,127.6
49Qwen Image2512· LMArena data1,123.9
50Wan 2.5 Preview· LMArena data1,117.4

Showing 1–50 of 86 results.

Data credits: LMArena, CC BY 4.0 (licence)