Skip to content

Benchmark

Arena ELO Text to Image - 3d Modeling

Higher scores are better.

Models measured
79
Scores recorded
86
Sources
1
IQ weight
1.29

Score distribution

1,145.6
Mean
1,144.3
Median
108.9
Std dev
1,015.4
10th percentile
1,295.7
90th percentile
861.9
Min
861.91,415
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.29 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 data1,415
2GPT-Image-2.6 SunburstMedium· LMArena data1,414.2
3GPT-Image-2.5 Flare· LMArena data1,383.1
4GPT Image 2Medium· LMArena data1,364.5
5MAI Image 2.7· LMArena data1,363.5
6GPT-Image-2.6 FlareMedium· LMArena data1,344.2
7Gemini Nano Banana 2.1· LMArena data1,328
8MAI-Image-2.6· LMArena data1,324.6
9Grok Imagine Image 2.0· LMArena data1,295.7
10Reve 2.1· LMArena data1,284.1
11Grok Imagine Image 2.020260801· LMArena data1,263.2
12Gemini 3.1 Flash Imagenano-banana-2, Web Search· LMArena data1,259.1
13Qwen Image 3.0 Pro· LMArena data1,252.2
14Mai Image 2.5· LMArena data1,251.4
15Seedream 5.0 Pro· LMArena data1,241.8
16Muse Image· LMArena data1,236.4
17Gemini 3 Pro Imagenano-banana-pro, 2K· LMArena data1,233.9
18Gemini 3.1 Flash-Lite Imagenano-banana-2-lite· LMArena data1,225.7
19Reve 2.0· LMArena data1,222.1
20GPT Image 1.5High Fidelity· LMArena data1,218
21Qwen Image 2.1· LMArena data1,216.5
22Gemini 3 Pro Imagenano-banana-pro· LMArena data1,210.5
23MiniMax H3-Image· LMArena data1,201.6
24Qwen Image 2.0 Pro2026-06-22· LMArena data1,195.2
25MAI-Image-2· LMArena data1,180.2
26Grok Imagine Image· LMArena data1,176.3
27Cosmos3-Super-Text2Image· LMArena data1,175.9
28Cosmos3-Super-Text2ImageAgentic· LMArena data1,174.2
29Ideogram V4Quality· LMArena data1,173.5
30UNI-1.1 Max· LMArena data1,173.4
31FLUX.2 [flex]· LMArena data1,173.3
32Hunyuan Image V3· LMArena data1,167.8
33UNI-1.1· LMArena data1,166.4
34Grok Imagine Image Pro· LMArena data1,165.9
35FLUX.2 [max]· LMArena data1,160.7
36Recraft V4.1 Utility Pro· LMArena data1,151.6
37Wan 2.6· LMArena data1,150.3
38FLUX.2 [dev]· LMArena data1,148.9
39Doubao Seedream 4-02K· LMArena data1,148.7
40Doubao Seedream 4.5· LMArena data1,145.7
41FLUX.2 [pro]· LMArena data1,145.5
42Reve v1.5· LMArena data1,145.1
43Krea 2 Turbo· LMArena data1,144.3
44Seedream 5.0 Lite· LMArena data1,142.5
45Gemini 2.5 Flash Imagenano-banana· LMArena data1,140.6
46Imagen 4 Ultragenerate-001· LMArena data1,138.9
47Doubao Seedream 4-0fal· LMArena data1,136.8
48Qwen Image2512· LMArena data1,135.4
49Wan 2.5 Preview· LMArena data1,134.9
50Krea v2 Medium· LMArena data1,133.3

Showing 1–50 of 86 results.

Data credits: LMArena, CC BY 4.0 (licence)