Skip to content

Benchmark

Arena ELO Text to Image - Art

Higher scores are better.

Models measured
79
Scores recorded
86
Sources
1
IQ weight
1.31

Score distribution

1,151.7
Mean
1,148.9
Median
110.2
Std dev
1,016.9
10th percentile
1,294.9
90th percentile
907.4
Min
907.41,420.8
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.31 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-2.6 SunburstMedium· LMArena data1,420.8
2GPT-Image-2.5 Sunburst· LMArena data1,414.7
3GPT-Image-2.5 Flare· LMArena data1,391.2
4MAI Image 2.7· LMArena data1,386.3
5GPT Image 2Medium· LMArena data1,374.2
6GPT-Image-2.6 FlareMedium· LMArena data1,373.7
7MAI-Image-2.6· LMArena data1,334.5
8Grok Imagine Image 2.0· LMArena data1,332.2
9Reve 2.1· LMArena data1,294.9
10Gemini Nano Banana 2.1· LMArena data1,284.7
11Mai Image 2.5· LMArena data1,276.2
12Grok Imagine Image 2.020260801· LMArena data1,272.5
13Seedream 5.0 Pro· LMArena data1,258.8
14Gemini 3.1 Flash-Lite Imagenano-banana-2-lite· LMArena data1,257.9
15Reve 2.0· LMArena data1,257.9
16Muse Image· LMArena data1,255.8
17Qwen Image 3.0 Pro· LMArena data1,252.6
18Gemini 3.1 Flash Imagenano-banana-2, Web Search· LMArena data1,239.6
19Qwen Image 2.1· LMArena data1,230
20GPT Image 1.5High Fidelity· LMArena data1,225.3
21Gemini 3 Pro Imagenano-banana-pro, 2K· LMArena data1,224.9
22Gemini 3 Pro Imagenano-banana-pro· LMArena data1,220.1
23Ideogram V4Quality· LMArena data1,200
24MiniMax H3-Image· LMArena data1,196.8
25UNI-1.1 Max· LMArena data1,190.1
26UNI-1.1· LMArena data1,188.9
27Cosmos3-Super-Text2ImageAgentic· LMArena data1,186.8
28Recraft V4.1 Utility Pro· LMArena data1,182.5
29Qwen Image 2.0 Pro2026-06-22· LMArena data1,178.6
30MAI-Image-2· LMArena data1,176.7
31Grok Imagine Image· LMArena data1,174.7
32Cosmos3-Super-Text2Image· LMArena data1,171.4
33FLUX.2 [max]· LMArena data1,168.1
34Doubao Seedream 4-02K· LMArena data1,164.2
35Grok Imagine Image Pro· LMArena data1,162.4
36Krea v2 Medium· LMArena data1,160.9
37Recraft V4.1 Pro· LMArena data1,160
38FLUX.2 [pro]· LMArena data1,159.2
39Wan 2.6· LMArena data1,158.5
40Doubao Seedream 4.5· LMArena data1,157.5
41Hunyuan Image V3· LMArena data1,156.2
42FLUX.2 [flex]· LMArena data1,155.8
43FLUX.2 [dev]· LMArena data1,148.9
44Gemini 2.5 Flash Imagenano-banana· LMArena data1,145.2
45Reve v1.5· LMArena data1,141.9
46Seedream 5.0 Lite· LMArena data1,141.7
47Krea 2 Turbo· LMArena data1,139
48Qwen Image2512· LMArena data1,130.8
49Recraft V4· LMArena data1,127.4
50Imagen 4 Ultragenerate-001· LMArena data1,125.9

Showing 1–50 of 86 results.

Data credits: LMArena, CC BY 4.0 (licence)