Candid Street Photography

Text-to-Image Photorealism

22 models were given the same prompt, and the community voted blind on which outputs looked best. How it works

Winning image by FLUX.2 [max] for Candid Street Photography

#1 — FLUX.2 [max]

Prompt

“A candid street photo of an elderly Japanese man repairing a red bicycle in light rain, reflections on wet pavement, shallow depth of field, 50mm lens, natural skin texture, imperfect framing, motion blur from passing cars, cinematic but realistic, no stylization.”

Challenge Rankings

22 models

#	Model	Price	¢/img	Elo
1	FLUX.2 [max] Black Forest Labs	$$	3¢	1227
2	Nano Banana Pro Google	$$$	6.7¢	1218
3	Seedream 4.5 ByteDance	$$	4¢	1216
4	Grok Imagine Image Pro xAI	$$$	7¢	1212
5	Seedream 5.0 Lite ByteDance	$$	3.5¢	1178
6	GPT Image 1.5 OpenAI	$	0.9¢	1174
7	Qwen Image 2512 Alibaba	$$	2¢	1173
8	GPT Image 1 Mini OpenAI	$	0.5¢	1171
9	ImagineArt 1.5 (Preview) Vyro AI	$$	3¢	1169
10	Imagen 4.0 Fast Generate 001 Google	$$	2¢	1168
11	Seedream 4.0 ByteDance	$$	3¢	1164
12	FLUX.2 [pro] Black Forest Labs	$$	1.5¢	1159
13	Grok Imagine Image xAI	$$	2¢	1156
14	Nano Banana Google	$$	3.9¢	1148
15	Nano Banana 2 Google	$$	2.2¢	1147
16	Z-Image Turbo Alibaba	$	0.5¢	1143
17	Stable Diffusion 3.5 Large Stability AI	$$$	6.5¢	1138
18	FLUX.2 [dev] Black Forest Labs	$$	1.2¢	1126
19	FLUX.2 [dev] Turbo fal	$	0.8¢	1124
20	FLUX.2 [flex] Black Forest Labs	$$$	6¢	1119
21	FLUX.2 [dev] Flash fal	$	0.5¢	1106
22	Imagen 4.0 Ultra Generate 001 Google	$$$	6¢	1078