llm leaderboard

Best AI for image generation

Human preference across generation and editing, rated blind, pairwise.

Aldena runs these models inside your team's projects. See what each one costs.

95 of 95 ranked models
Ranked models
rankmodelvendorcompositebenchmarkseloeloelo
1
GPT Image 2.5 Sunburstopenai/gpt-image-2.5-sunburst
OpenAI81.93 of 3 benchmarks142815201525
2
GPT Image 2.5 Flareopenai/gpt-image-2.5-flare
OpenAI80.63 of 3 benchmarks140014851476
3
GPT Image 2 (medium)openai/gpt-image-2:medium
OpenAI78.93 of 3 benchmarks138214611455
4
GPT Image 2.6 Flare (medium)openai/gpt-image-2.6-flare:medium
OpenAI78.43 of 3 benchmarks138314601439
5
GPT Image 2.6 Sunburst (medium)openai/gpt-image-2.6-sunburst:medium
OpenAI76.22 of 3 benchmarks14061477
6
Nano Banana 2.1google/gemini-nano-banana-2.1
Google75.73 of 3 benchmarks132814271431
7
MAI-Image-2.6microsoft/mai-image-2.6
Microsoft AI73.22 of 3 benchmarks13301428
8
Grok Imagine Image 2.0x-ai/grok-imagine-image-2.0
SpaceXAI73.02 of 3 benchmarks13191437
9
Muse Imagemeta/muse-image
Meta72.73 of 3 benchmarks126614001399
10
Krea 3 Ultrakrea/krea-3-ultra
Krea71.62 of 3 benchmarks14111433
11
Seedream 5.0 Probytedance-seed/seedream-5-0-pro
ByteDance Seed71.63 of 3 benchmarks124913921412
12
Nano Banana 2 (Gemini 3.1 Flash Image)google/gemini-3.1-flash-image
Google70.33 of 3 benchmarks125713871370
13
MAI-Image-2.5microsoft/mai-image-2.5
Microsoft AI69.32 of 3 benchmarks12531400
14
Nano Banana Pro (Gemini 3 Pro Image) (2K)google/gemini-3-pro-image:2k
Google68.83 of 3 benchmarks124613911364
15
Nano Banana Pro (Gemini 3 Pro Image Preview)google/gemini-3-pro-image-preview
Google67.73 of 3 benchmarks123113851366
16
Reve 2.0reve/reve-2.0
Reve67.63 of 3 benchmarks126813591345
17
Reve 2.1reve/reve-2.1
Reve67.52 of 3 benchmarks13021373
18
GPT Image 1.5 High Fidelityopenai/gpt-image-1.5-high-fidelity
OpenAI65.53 of 3 benchmarks123613671342
19
Chatgpt Image High Fidelityopenai/chatgpt-image-high-fidelity
OpenAI65.12 of 3 benchmarks13871351
20
Qwen Image 3.0 Proqwen/qwen-image-3.0-pro
Qwen64.71 of 3 benchmarks1254
21
Grok Imagine Image Qualityx-ai/grok-imagine-image-quality
SpaceXAI62.91 of 3 benchmarks1391
22
Uni-1.1 (max)luma/uni-1.1:max
Luma AI62.53 of 3 benchmarks118813331308
23
Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)google/gemini-3.1-flash-lite-image
Google61.93 of 3 benchmarks124913131289
24
Ideogram 4.0 Qualityideogram/ideogram-4.0-quality
Ideogram61.81 of 3 benchmarks1199
25
Ideogram 4.5ideogram/ideogram-4.5
Ideogram60.82 of 3 benchmarks13461333
26
Mai Image 2microsoft/mai-image-2
Microsoft60.51 of 3 benchmarks1183
27
Qwen Image 2.0 Proqwen/qwen-image-2.0-pro
Qwen60.22 of 3 benchmarks11901303
28
Grok Imagine Imagex-ai/grok-imagine-image
SpaceXAI60.02 of 3 benchmarks11681328
29
Recraft V4.1 Utility Prorecraft/recraft-v4.1-utility-pro
Recraft59.71 of 3 benchmarks1169
30
Uni-1.1luma/uni-1.1
Luma AI59.53 of 3 benchmarks118213161287
31
Cosmos3 Super Text2imagenvidia/cosmos3-super-text2image
NVIDIA58.81 of 3 benchmarks1164
32
Grok Imagine Image Prox-ai/grok-imagine-image-pro
SpaceXAI58.01 of 3 benchmarks1160
33
Reve v1.5reve/reve-v1.5
Reve57.11 of 3 benchmarks1154
34
Seedream 4.5bytedance-seed/seedream-4.5
ByteDance Seed57.03 of 3 benchmarks114813031290
35
Hunyuan Image 3.0tencent/hunyuan-image-3.0
Tencent56.31 of 3 benchmarks1152
36
Imagen Ultra 4.0 Generate 001google/imagen-ultra-4.0-generate-001
Google55.01 of 3 benchmarks1146
37
Wan2.6 T2Ialibaba/wan2.6-t2i
Alibaba53.81 of 3 benchmarks1138
38
Hunyuan Image 3.0 Instructtencent/hunyuan-image-3.0-instruct
Tencent53.71 of 3 benchmarks1301
39
Recraft V4.1 Prorecraft/recraft-v4.1-pro
Recraft52.91 of 3 benchmarks1132
40
FLUX.2 Maxblack-forest-labs/flux.2-max
Black Forest Labs52.83 of 3 benchmarks116112621250
41
Seedream 5.0 Litebytedance-seed/seedream-5-0-lite
ByteDance Seed52.63 of 3 benchmarks114112931266
42
Krea 2 Mediumkrea/krea-2-medium
Krea52.51 of 3 benchmarks1128
43
Imagen 4.0 Generate 001google/imagen-4.0-generate-001
Google52.11 of 3 benchmarks1128
44
Qwen Image 2512qwen/qwen-image-2512
Qwen51.61 of 3 benchmarks1124
45
Wan2.7 Image Proalibaba/wan2.7-image-pro
Alibaba51.43 of 3 benchmarks110313021290
46
Recraft V4recraft/recraft-v4
Recraft50.81 of 3 benchmarks1116
47
Nano Banana (Gemini 2.5 Flash Image)google/gemini-2.5-flash-image
Google50.83 of 3 benchmarks114812921234
48
FLUX.2 Problack-forest-labs/flux.2-pro
Black Forest Labs50.23 of 3 benchmarks115312481240
49
Reve v1.1reve/reve-v1.1
Reve50.22 of 3 benchmarks12611261
50
Krea 2 Turbokrea/krea-2-turbo
Krea50.01 of 3 benchmarks1116
51
Kling Image O1kwaivgi/kling-image-o1
Kling AI49.22 of 3 benchmarks12501253
52
Wan2.7 Imagealibaba/wan2.7-image
Alibaba49.13 of 3 benchmarks110013001280
53
Hidream O1 Imagehidream/hidream-o1-image
HiDream49.11 of 3 benchmarks1115
54
Krea 2 Largekrea/krea-2-large
Krea48.71 of 3 benchmarks1114
55
Qwen Image Editqwen/qwen-image-edit
Qwen48.51 of 3 benchmarks1241
56
FLUX.2 Flexblack-forest-labs/flux.2-flex
Black Forest Labs48.03 of 3 benchmarks115612241234
57
Recraft V4.1 Flashrecraft/recraft-v4.1-flash
Recraft46.61 of 3 benchmarks1098
58
Mai Image 1microsoft/mai-image-1
Microsoft46.21 of 3 benchmarks1093
59
Seedream 4 (2K)bytedance/seedream-4:2k
ByteDance46.03 of 3 benchmarks113812671198
60
Z Image Turboalibaba/z-image-turbo
Alibaba45.71 of 3 benchmarks1082
61
Seedream 3bytedance/seedream-3
ByteDance45.31 of 3 benchmarks1081
62
Reve v1reve/reve-v1
Reve45.12 of 3 benchmarks12351224
63
Qwen Image Prompt Extendqwen/qwen-image-prompt-extend
Qwen44.01 of 3 benchmarks1061
64
Wan2.6 Imagealibaba/wan2.6-image
Alibaba43.62 of 3 benchmarks12311220
65
Flux 2 Devblack-forest-labs/flux-2-dev
Black Forest Labs43.43 of 3 benchmarks114512241202
66
Reve v1.1 Fastreve/reve-v1.1-fast
Reve43.41 of 3 benchmarks1207
67
Qwen Imageqwen/qwen-image
Qwen43.21 of 3 benchmarks1057
68
Imagen 3.0 Generate 002google/imagen-3.0-generate-002
Google42.81 of 3 benchmarks1057
69
Reve Edit Fastreve/reve-edit-fast
Reve42.81 of 3 benchmarks1198
70
Ideogram v3 Qualityideogram/ideogram-v3-quality
Ideogram42.41 of 3 benchmarks1049
71
Photonluma/photon
Luma AI41.91 of 3 benchmarks1036
72
Qwen Image Edit 2511qwen/qwen-image-edit-2511
Qwen41.22 of 3 benchmarks12331173
73
Runway Gen4runway/runway-gen4
Runway41.11 of 3 benchmarks1024
74
Recraft V3recraft/recraft-v3
Recraft40.71 of 3 benchmarks1022
75
Flux 1.1 Problack-forest-labs/flux-1.1-pro
Black Forest Labs40.31 of 3 benchmarks1016
76
Ideogram v2ideogram/ideogram-v2
Ideogram39.81 of 3 benchmarks1013
77
Seedream 4 High Res Falbytedance/seedream-4-high-res-fal
ByteDance39.73 of 3 benchmarks111312161209
78
Lucid Originleonardo/lucid-origin
Leonardo AI39.41 of 3 benchmarks1013
79
GLM Imagez-ai/glm-image
Z.ai39.01 of 3 benchmarks1012
80
Flux 2 Klein 9Bblack-forest-labs/flux-2-klein-9b
Black Forest Labs38.93 of 3 benchmarks107012241211
81
Flux 1 Devblack-forest-labs/flux-1-dev
Black Forest Labs38.11 of 3 benchmarks969
82
Seedream 4 Falbytedance/seedream-4-fal
ByteDance37.93 of 3 benchmarks111612091158
83
DALL E 3openai/dall-e-3
OpenAI37.71 of 3 benchmarks968
84
Wan2.5 I2I Previewalibaba/wan2.5-i2i-preview
Alibaba37.13 of 3 benchmarks111711801165
85
Stable Diffusion v35 Largestability-ai/stable-diffusion-v35-large
Stability AI36.91 of 3 benchmarks938
86
Step1x Editstepfun/step1x-edit
StepFun36.51 of 3 benchmarks996
87
GPT Image 1openai/gpt-image-1
OpenAI33.83 of 3 benchmarks111511391126
88
FLUX.2 Klein 4Bblack-forest-labs/flux.2-klein-4b
Black Forest Labs32.53 of 3 benchmarks103011871166
89
GPT Image 1 Miniopenai/gpt-image-1-mini
OpenAI32.03 of 3 benchmarks110911241111
90
Flux 1 Kontext (max)black-forest-labs/flux-1-kontext:max
Black Forest Labs31.43 of 3 benchmarks107511821062
91
Flux 1 Kontext Problack-forest-labs/flux-1-kontext-pro
Black Forest Labs30.43 of 3 benchmarks105911771063
92
Seededit 3.0bytedance/seededit-3.0
ByteDance30.12 of 3 benchmarks11391042
93
Bagelbytedance/bagel
ByteDance27.82 of 3 benchmarks8971027
94
Gemini 2.0 Flash Preview Image Generationgoogle/gemini-2.0-flash-preview-image-generation
Google24.73 of 3 benchmarks97610821058
95
Flux 1 Kontext Devblack-forest-labs/flux-1-kontext-dev
Black Forest Labs24.43 of 3 benchmarks94111501036
How this ranks

Every benchmark value becomes a percentile among the models that have it, so accuracy scores, Elo ratings and word error rates compare without hand-tuned scaling. Metrics where lower is better are inverted first. Raw values are never summed or averaged across benchmarks. A model's mean percentile is then shrunk toward the mean of the models that were broadly benchmarked, so a model tested twice cannot outrank a broadly tested one on two lucky results. Turning a data source off runs that same ranking code again in your browser over the sources you left on.

A model scored on fewer than 2 of the 3 ranked benchmarks in this category still ranks here, on the benchmarks it does have, and its row carries a partial coverage mark. On an equal score it sits under the model that earned the same number across more of the board.

Data sources

Turn a source off to drop every benchmark it feeds and re-rank the board in your browser. Your choice follows you across the leaderboard pages.

  • LMArenaCC BY 4.0

    Arena ratings by LMArena, from the public leaderboard dataset.