How the score works

  • Prompt adherence25%
  • Style fidelity13%
  • Quality & detail15%
  • Realism & anatomy15%
  • Composition & framing12%
  • Look & lighting10%
  • Text rendering10%

A criterion that does not apply is skipped and its weight is shared by the rest, so a prompt that asks for no text in the image never costs a model points on text rendering.

Overall ranking

# Model imagemodels.ai score Prompt adherence Style fidelity Quality & detail Realism & anatomy Composition & framing Look & lighting Text rendering Best at
1 Nano Banana 2.1 41 prompts 4.52 4.46 4.67 4.55 4.44 4.55 4.50 4.58 Typography & Posters 4.73
2 Qwen Image 2.1 27 prompts 4.51 4.44 4.69 4.61 4.48 4.39 4.43 4.89 Logo & Graphic Design 4.69
3 Nano Banana 2.1 Advanced 76 prompts 4.45 4.43 4.54 4.49 4.41 4.44 4.40 4.52 Typography & Posters 4.56
4 Alibaba Qwen Image 3.0 Pro 76 prompts 4.39 4.38 4.47 4.49 4.20 4.42 4.38 4.52 Typography & Posters 4.58
5 Qwen Image 3.0 47 prompts 4.34 4.24 4.48 4.45 4.24 4.34 4.33 4.71 Typography & Posters 4.57
6 GPT Image 2.5 Flare 76 prompts 4.32 4.24 4.32 4.45 4.19 4.39 4.36 4.55 Infographics & UI 4.60
7 GPT Image 2.5 Sunburst 76 prompts 4.28 4.14 4.29 4.43 4.14 4.38 4.34 4.64 Illustration & Anime 4.61
8 GPT Image 2 76 prompts 4.22 4.15 4.03 4.40 4.19 4.34 4.11 4.86 Illustration & Anime 4.60
9 Grok Imagine Image V2 76 prompts 4.15 4.01 4.16 4.28 4.09 4.22 4.20 4.48 Infographics & UI 4.58
10 Nano Banana 2 76 prompts 4.10 3.96 4.00 4.26 4.11 4.22 4.07 4.52 Sci-Fi & Fantasy 4.47
11 Seedream 5.0 Pro 76 prompts 4.10 3.98 3.97 4.26 4.05 4.15 4.15 4.75 Counting & Composition 4.68
12 GPT Image 1.5 76 prompts 4.05 3.90 4.05 4.24 3.95 4.17 4.13 3.98 3D & Renders 4.30
13 Nano Banana Pro 76 prompts 4.04 3.98 3.88 4.16 3.99 4.19 3.99 4.57 Counting & Composition 4.43
14 FLUX.2 Flex 76 prompts 4.04 3.84 4.03 4.24 3.96 4.17 4.16 4.25 Typography & Posters 4.35
15 Z Image Pro Uncensored 1 prompt 3.99 3.50 4.50 4.50 3.50 4.00 4.50 — Animals & Wildlife 3.99
16 Nano Banana 76 prompts 3.98 3.86 3.86 4.16 3.94 4.14 3.92 4.24 Glass, Water & Reflections 4.37
17 Nano Banana 2 Lite 76 prompts 3.97 3.86 3.93 4.09 3.92 4.09 3.97 4.07 Sci-Fi & Fantasy 4.42
18 Seedream 5.0 Lite 76 prompts 3.97 3.86 3.80 4.13 3.89 4.02 4.01 4.73 Signs & World Scripts 4.37
19 Seedream 4.0 76 prompts 3.96 3.71 3.89 4.20 3.88 4.16 4.03 4.43 Food & Drink 4.30
20 FLUX.2 Pro 76 prompts 3.95 3.69 4.01 4.16 3.89 4.05 4.07 4.27 Signs & World Scripts 4.29
21 FLUX.2 Dev 76 prompts 3.93 3.66 3.97 4.18 3.82 4.13 4.11 4.05 Product & Packshot 4.29
22 Qwen Image Max Uncensored 76 prompts 3.90 3.27 4.30 4.26 3.72 4.15 4.41 3.61 Typography & Posters 4.31
23 Grok Imagine Image 76 prompts 3.89 3.72 3.89 4.05 3.74 4.00 4.03 4.43 Sci-Fi & Fantasy 4.17
24 FLUX.2 Klein 9B 76 prompts 3.85 3.63 3.82 4.08 3.80 4.04 4.01 3.29 Fashion & Editorial 4.35
25 FLUX.2 Dev Turbo 76 prompts 3.85 3.60 3.99 4.06 3.80 4.01 3.95 3.68 Fashion & Editorial 4.24
26 ERNIE Image 76 prompts 3.82 3.55 3.68 4.15 3.79 4.07 3.90 3.98 Signs & World Scripts 4.18
27 Seedream 4.5 76 prompts 3.81 3.54 3.72 4.07 3.76 3.99 3.86 4.62 Signs & World Scripts 4.25
28 Microsoft Lens 76 prompts 3.74 3.44 3.64 4.09 3.68 4.01 3.81 3.90 Glass, Water & Reflections 4.07
29 P-Image Ideogram 76 prompts 3.72 3.46 3.68 3.86 3.80 3.80 3.79 4.38 Signs & World Scripts 4.41
30 ERNIE Image Turbo 76 prompts 3.67 3.27 3.56 4.07 3.70 3.93 3.88 3.77 Signs & World Scripts 4.12
31 SenseNova U1 8B 76 prompts 3.63 3.39 3.55 3.84 3.57 3.88 3.72 3.90 Portraits & Faces 4.08
32 LongCat Image 76 prompts 3.63 3.26 3.52 3.97 3.64 3.82 3.87 3.67 Product & Packshot 4.12
33 FLUX.1 Krea Dev 76 prompts 3.63 3.21 3.63 4.03 3.70 3.89 3.84 2.58 Portraits & Faces 4.08
34 Qwen Image 76 prompts 3.62 3.36 3.59 3.80 3.65 3.77 3.81 3.57 Portraits & Faces 3.94
35 FLUX.2 Klein 4B 76 prompts 3.61 3.18 3.63 4.00 3.64 3.82 3.92 3.00 Fashion & Editorial 4.06
36 FLUX 1.1 Pro 76 prompts 3.61 3.09 3.57 4.03 3.69 3.97 3.89 3.07 Portraits & Faces 3.90
37 HiDream I1 Dev 76 prompts 3.59 3.09 3.55 4.04 3.69 3.91 3.83 2.75 Fashion & Editorial 4.13
38 Z-Image Turbo 76 prompts 3.58 3.24 3.54 3.82 3.70 3.70 3.76 3.63 Portraits & Faces 4.05
39 FLUX 1.1 Pro Ultra 76 prompts 3.58 3.11 3.55 3.96 3.68 3.82 3.88 2.98 Landscape & Nature 3.86
40 HiDream I1 Fast 76 prompts 3.54 3.00 3.47 4.01 3.56 3.95 3.85 2.73 Portraits & Faces 3.92
41 HiDream I1 Full 76 prompts 3.47 2.97 3.41 3.83 3.55 3.78 3.84 2.93 Glass, Water & Reflections 3.76
42 GLM Image 76 prompts 3.47 3.04 3.36 3.94 3.52 3.78 3.68 3.14 Typography & Posters 3.97
43 Wan Image Small 76 prompts 3.45 2.90 3.44 3.89 3.63 3.75 3.74 2.25 Fashion & Editorial 3.92
44 FLUX.1 Schnell 76 prompts 3.40 2.91 3.45 3.80 3.36 3.69 3.79 2.66 Glass, Water & Reflections 3.99
45 FLUX.1 Dev 76 prompts 3.39 2.88 3.38 3.76 3.49 3.73 3.67 2.55 Portraits & Faces 3.90
46 P-Image 76 prompts 3.37 2.79 3.30 3.89 3.64 3.64 3.52 2.50 Landscape & Nature 3.83
47 Hunyuan FLUX SRPO 76 prompts 3.13 2.55 3.14 3.47 3.28 3.51 3.51 2.34 Portraits & Faces 3.85

Category winners

3D & Renders Nano Banana 2.1
4.63
Animals & Wildlife Nano Banana 2.1
4.49
Architecture & Interiors Alibaba Qwen Image 3.0 Pro
4.55
Cars & Vehicles GPT Image 2.5 Flare
4.43
Counting & Composition Seedream 5.0 Pro
4.68
Fashion & Editorial Nano Banana 2.1 Advanced
4.45
Food & Drink Nano Banana 2.1 Advanced
4.54
Glass, Water & Reflections Alibaba Qwen Image 3.0 Pro
4.49
Hands & Gestures Nano Banana 2.1 Advanced
4.42
Illustration & Anime GPT Image 2.5 Sunburst
4.61
Infographics & UI GPT Image 2.5 Sunburst
4.60
Landscape & Nature GPT Image 2.5 Flare
4.58
Logo & Graphic Design Qwen Image 2.1
4.69
Macro & Close-ups Nano Banana 2.1 Advanced
4.50
Portraits & Faces Nano Banana 2.1 Advanced
4.45
Product & Packshot Qwen Image 2.1
4.66
Sci-Fi & Fantasy Grok Imagine Image V2
4.55
Signs & World Scripts GPT Image 2.5 Sunburst
4.61
Typography & Posters Nano Banana 2.1
4.73

Cite this leaderboard

Wiro AI (2026). Image Models benchmark: AI image model leaderboard, updated 30 October 2026 [Data set]. https://imagemodels.ai/leaderboard/. Licensed under CC BY 4.0.

CC BY 4.0. BibTeX, both CSV files and how every score is made are on the methodology page.