Text-to-Image Arena🏆Overall

View overall rankings across text to image AI models.

Jun 5, 2026
5,391,418 votes
20 labs
Lab Rank
Model Score
Rank Spread
1
OpenAI
gpt-image-2 (medium) · Proprietary
1385±6
1
11
2
Reve
reve-2.0 · Proprietary
1273±9
2
23
3
Google
gemini-3.1-flash-image-preview (nano-banana-2) [web-search] · Proprietary
1269±4
3
23
4
Microsoft AI
mai-image-2.5 · Proprietary
1253±6
4
45
5
xAI
grok-imagine-image-quality · Proprietary
1234±6
7
68
6
Ideogram
Ideogram
ideogram-4.0-quality
1204±10
9
910
7
Luma AI
uni-1.1-max · Proprietary
1191±7
10
911
8
Recraft
Recraft
recraft-v4.1-utility-pro · Proprietary
1169±11
14
1120
9
Alibaba
qwen-image-2.0-pro-2026-04-22 · Proprietary
1169±8
15
1217
10
Flux
Black Forest Labs
flux-2-max · Proprietary
1163±4
16
1419
11
Tencent
Tencent
hunyuan-image-3.0
1151±3
22
1825
12
Bytedance
Bytedance
seedream-4.5 · Proprietary
1145±3
25
2126
13
HiDream
hidream-o1-image
1124±7
32
2737
14
Krea
krea-2-medium · Proprietary
1121±11
33
2741
15
Nvidia
Cosmos3-Super-Text2Image
1062±10
49
4755
16
Pruna
p-image · Proprietary
1033±4
57
5658
17
Runway
runway-gen4 · Proprietary
1025±5
59
5860
18
Leonardo AI
lucid-origin · Proprietary
1014±3
62
6164
19
Z.ai
glm-image
1011±9
64
6064
20
Stability
Stability AI
stable-diffusion-v35-large
938±4
69
6869

Remove Style Control Leaderboard Plots

Confidence Intervals on Model Strength (via Bootstrapping)

Battle Count for Each Combination of Models (without Ties)

Fraction of Model A Wins for All Non-tied A vs. B Battles

Average Win Rate Against All Other Models (Uniform Sampling and No Ties)