Text to Image · Photorealistic & Cinematic Imagery

Best AI for photorealistic images: AI Image Arena ranked

Nov 17, 2026 · 2,269,818 votes · 86 models

GPT-Image-2.5 Sunburst by OpenAI leads with 1434 points, ahead of GPT-Image-2.5 Flare (1417) and GPT-Image-2.6 Sunburst (Medium) (1411).

Lab
License
RankModelScoreSpreadVotes
1
GPT-Image-2.5 SunburstOpenAI · Proprietary
1434 ±8
123,046
2
GPT-Image-2.5 FlareOpenAI · Proprietary
1417 ±7
2-322,431
3
GPT-Image-2.6 Sunburst (Medium)OpenAI · Proprietary
1411 ±14
2-52,331
4
MAI-Image-2.7Microsoft AI · Proprietary
1389 ±20
3-61,032
5
GPT-Image-2.6 Flare (Medium)OpenAI · Proprietary
1388 ±14
3-62,234
6
GPT-Image-2 (Medium)OpenAI · Proprietary
1379 ±5
4-640,767
7
Gemini Nano Banana 2.1Google · Proprietary
1337 ±15
7-91,878
8
MAI-Image-2.6Microsoft AI · Proprietary
1331 ±7
7-914,773
9
Grok Imagine Image 2.0xAI · Proprietary
1314 ±11
7-103,408
10
Reve 2.1Reve · Proprietary
1304 ±11
9-113,156
11
1292 ±7
10-1213,659
12
Reve 2.0Reve · Proprietary
1277 ±9
11-156,291
13
Muse ImageMeta · Proprietary
1274 ±6
12-1522,368
14
Seedream 5.0 ProByteDance · Proprietary
1273 ±5
12-1542,507
15
1260 ±4
15-1976,645
16
1260 ±5
15-1931,682
17
Qwen-Image-3.0-ProAlibaba · Proprietary
1259 ±10
12-204,124
18
1259 ±8
15-207,057
19
MAI-Image-2.5Microsoft AI · Proprietary
1251 ±6
15-2122,544
20
GPT-Image-1.5 (High Fidelity)OpenAI · Proprietary
1247 ±5
17-2166,877
21
1240 ±6
19-2237,479
22
Qwen-Image-2.1Alibaba · qwen-research
1232 ±9
21-236,186
23
MiniMax H3-ImageMiniMax · minimax-h3-image-community-license-agreement
1208 ±17
22-281,163
24
Qwen-Image-2.0-Pro (2026-06-22)Alibaba · Proprietary
1207 ±9
23-275,076
25
MAI-Image-2Microsoft AI · Proprietary
1203 ±7
23-2820,427
26
Ideogram 4.0 (Quality)Ideogram · Ideogram Open Model
1200 ±5
23-2831,687
27
UNI-1.1 MaxLuma AI · Proprietary
1188 ±9
24-305,733
28
UNI-1.1Luma AI · Proprietary
1182 ±6
27-3024,174
29
Recraft V4.1 Utility ProRecraft · Proprietary
1182 ±18
23-35997
30
Grok Imagine ImagexAI · Proprietary
1176 ±4
27-32118,230
31
Grok Imagine Image ProxAI · Proprietary
1169 ±5
29-3540,108
32
Seedream 4.5ByteDance · Proprietary
1166 ±4
30-36126,231
33
1166 ±10
29-413,871
34
FLUX.2 [max]Black Forest Labs · Proprietary
1161 ±5
30-4151,762
35
Seedream 5.0 LiteByteDance · Proprietary
1158 ±4
32-4259,612
36
1154 ±3
33-43412,919
37
HunyuanImage 3.0Tencent · tencent-hunyuan-community
1153 ±5
33-4477,658
38
Imagen 4 Ultra (generate-001)Google · Proprietary
1153 ±6
33-45185,868
39
FLUX.2 [pro]Black Forest Labs · Proprietary
1152 ±4
35-4591,155
40
Seedream 4.0 (2K)ByteDance · Proprietary
1150 ±9
33-465,806
41
Cosmos3-Super-Text2ImageNVIDIA · OpenMDW-1.1
1149 ±9
33-466,647
42
Reve v1.5Reve · Proprietary
1147 ±6
36-4615,588
43
Recraft V4.1 ProRecraft · Proprietary
1147 ±17
30-511,124
44
FLUX.2 [flex]Black Forest Labs · Proprietary
1146 ±5
37-4665,120
45
FLUX.2 [dev]Black Forest Labs · flux-non-commercial-license
1143 ±5
38-4831,700
46
Krea 2 MediumKrea · Proprietary
1137 ±6
40-5129,738
47
Imagen 4 (generate-001)Google · Proprietary
1134 ±4
44-51248,320
48
Qwen-Image (2512)Alibaba · Apache 2.0
1133 ±5
44-5340,367
49
Wan 2.6Alibaba · Proprietary
1131 ±4
45-53101,219
50
Seedream 4.0 (High Res, fal)ByteDance · Proprietary
1130 ±5
45-5377,110
51
Krea 2 LargeKrea · Proprietary
1126 ±6
45-5825,251
52
Wan 2.7 Image ProAlibaba · Proprietary
1122 ±7
48-6011,773
53
Recraft V4Recraft · Proprietary
1120 ±4
51-6062,710
54
Seedream 4.0 (fal)ByteDance · Proprietary
1120 ±9
48-615,585
55
Krea 2 TurboKrea · krea-2-community-license
1118 ±6
51-6123,840
56
Wan 2.7 ImageAlibaba · Proprietary
1116 ±7
51-6312,273
57
Z-Image-TurboAlibaba · Apache 2.0
1116 ±7
51-6310,195
58
GPT-Image-1OpenAI · Proprietary
1114 ±4
52-62118,148
59
GPT-Image-1 miniOpenAI · Proprietary
1114 ±5
52-6369,804
60
Recraft V4.1 FlashRecraft · Proprietary
1113 ±11
51-643,424
61
HiDream-O1-ImageHiDream · MIT
1108 ±5
54-6429,143
62
MAI-Image-1Microsoft AI · Proprietary
1104 ±5
57-6440,361
63
Seedream 3.0ByteDance · Proprietary
1103 ±7
56-6517,149
64
Wan 2.5 PreviewAlibaba · Proprietary
1100 ±4
60-65117,334
65
FLUX.2 [klein] 9BBlack Forest Labs · flux-non-commercial-license
1092 ±5
63-6661,973
66
FLUX.1 Kontext [max]Black Forest Labs · Proprietary
1084 ±5
65-6630,177
67
Imagen 3 (generate-002)Google · Proprietary
1074 ±5
67-69159,135
68
P-ImageUnknown lab · Proprietary
1072 ±5
67-6944,606
69
FLUX.1 Kontext [pro]Black Forest Labs · Proprietary
1067 ±5
67-70158,964
70
Qwen-ImageAlibaba · Apache 2.0
1063 ±4
69-7142,836
71
Recraft V3Recraft · Proprietary
1056 ±5
70-7480,410
72
Ideogram 3.0 (Quality)Ideogram · Proprietary
1051 ±5
71-7651,674
73
Qwen-Image (Prompt Extend)Alibaba · Apache 2.0
1048 ±4
71-76334,776
74
FLUX.2 [klein] 4BBlack Forest Labs · Apache 2.0
1047 ±5
71-7662,685
75
FLUX1.1 [pro]Black Forest Labs · Proprietary
1044 ±5
72-7826,892
76
Runway Gen-4Runway · Proprietary
1040 ±6
72-7823,370
77
Lucid OriginLeonardo AI · Proprietary
1035 ±5
75-78132,040
78
Luma PhotonLuma AI · Proprietary
1035 ±5
75-7856,133
79
Ideogram 2.0Ideogram · Proprietary
1020 ±6
7927,531
80
GLM-ImageZ.ai · MIT
1000 ±13
80-812,035
81
FLUX.1 [dev] (FP8)Black Forest Labs · Open
990 ±6
80-8119,407
82
963 ±5
82122,526
83
DALL-E 3OpenAI · Proprietary
941 ±6
83-85104,171
84
FLUX.1 Kontext [dev]Black Forest Labs · flux-1-dev-non-commercial-license
937 ±5
83-85104,553
85
Stable Diffusion 3.5 LargeUnknown lab · Open
928 ±7
83-869,405
86
BAGELByteDance · Apache 2.0
922 ±9
85-865,461

Score: Bradley-Terry rating - the bar shows it on one scale for the whole board, the thin line is the 95% confidence interval (± half of it next to the number). Spread: every rank the model could hold given that interval - models whose lines overlap are statistically tied.

Data: LMArena (lmarena-ai/leaderboard-dataset), CC BY 4.0. Published Nov 17, 2026. Flashlook mirrors the rankings unchanged - about the data.

For fashion & e-commerce brands

Photorealism is the baseline for fashion imagery: skin, fabric and light have to read as a real photograph. For product pages you also need the garment to match the real item exactly - which a prompt-only image can't guarantee.

Shoot your products on-model

Confidence intervals on model strength

Dot = score, line = 95% interval. Hover a dot for the numbers.

GPT-Image-2.5 Sunburst
GPT-Image-2.5 Flare
GPT-Image-2.6 Sunburst (Medium)
MAI-Image-2.7
GPT-Image-2.6 Flare (Medium)
GPT-Image-2 (Medium)
Gemini Nano Banana 2.1
MAI-Image-2.6
Grok Imagine Image 2.0
Reve 2.1
Grok Imagine Image 2.0 (20260801)
Reve 2.0
Muse Image
Seedream 5.0 Pro
Gemini 3 Pro Image (nano-banana-pro, 2K)
Gemini 3.1 Flash Image (nano-banana-2, Web Search)
Qwen-Image-3.0-Pro
Gemini 3.1 Flash-Lite Image (nano-banana-2-lite)
MAI-Image-2.5
GPT-Image-1.5 (High Fidelity)
Gemini 3 Pro Image Preview (nano-banana-pro)
Qwen-Image-2.1
MiniMax H3-Image
Qwen-Image-2.0-Pro (2026-06-22)
MAI-Image-2
Ideogram 4.0 (Quality)
UNI-1.1 Max
UNI-1.1
Recraft V4.1 Utility Pro
Grok Imagine Image
Grok Imagine Image Pro
Seedream 4.5
Cosmos3-Super-Text2Image (Agentic)
FLUX.2 [max]
Seedream 5.0 Lite
Gemini 2.5 Flash Image Preview (nano-banana)
HunyuanImage 3.0
Imagen 4 Ultra (generate-001)
FLUX.2 [pro]
Seedream 4.0 (2K)
110011501200125013001350140014501500

Top 40 shown; the table lists all 86.