Arena Over-the-top cartoon caricature
Image Editing
13 models were given the same image and edit instruction, and the community voted blind on which outputs looked best.
How it works
#1 — Nano Banana 2
Source Image
Over-the-top cartoon caricature
Edit instruction
“ Create a caricature of me and my job. Make it exaggerated and humorous, incorporating my profession as a tv show anchor and my love for dogs and hockey. ”
Show all
Challenge Rankings
13 models
#
Model
Price
¢/img
Speed
Elo
1
$$
2.2¢
1295
2
$$$
6.7¢
1252
3
$
0.9¢
1243
4
$$
3.5¢
1229
5
$$
4¢
1215
6
$$
3¢
1202
7
$$
1.5¢
1201
8
$$
2¢
1197
9
$$$
7¢
1194
10
$$
3¢
1192
11
$$
3.9¢
1181
12
$$
4¢
1138
13
$$
3¢
1133
Google's Nano Banana 2 leads the challenge with 1295 Elo and a 75% win rate, holding a 49-point lead over the next model despite being three times cheaper and faster than the Nano Banana Pro. GPT Image 1.5 maintains a competitive 1243 Elo and 70% win rate while operating as the most cost-effective top-tier model at $0.009 per image.
4 models waiting for enough speed data
Competitors
13 models, ranked by Elo
Nano Banana 2
3 attempts · best result used for ranking
1295
Elo
80%W
·
10%L
·
10%T
1167
Elo
75%W
·
25%L
·
0%T
1083
Elo
50%W
·
50%L
·
0%T
Nano Banana Pro
3 attempts · best result used for ranking
1252
Elo
75%W
·
20%L
·
5%T
1193
Elo
60%W
·
40%L
·
0%T
1114
Elo
22%W
·
67%L
·
11%T
GPT Image 1.5
3 attempts · best result used for ranking
1243
Elo
75%W
·
17%L
·
8%T
1188
Elo
67%W
·
11%L
·
22%T
1158
Elo
44%W
·
44%L
·
12%T
Seedream 5.0 Lite
3 attempts · best result used for ranking
1229
Elo
50%W
·
25%L
·
25%T
1185
Elo
60%W
·
40%L
·
0%T
1039
Elo
25%W
·
75%L
·
0%T
Seedream 4.5
3 attempts · best result used for ranking
1215
Elo
68%W
·
32%L
·
0%T
1157
Elo
38%W
·
46%L
·
16%T
1147
Elo
30%W
·
50%L
·
20%T
Seedream 4.0
3 attempts · best result used for ranking
1202
Elo
50%W
·
50%L
·
0%T
1089
Elo
29%W
·
71%L
·
0%T
992
Elo
0%W
·
100%L
·
0%T
FLUX.2 [pro]
3 attempts · best result used for ranking
1201
Elo
63%W
·
31%L
·
6%T
1164
Elo
63%W
·
38%L
·
-1%T
1100
Elo
29%W
·
57%L
·
14%T
Grok Imagine Image
3 attempts · best result used for ranking
1197
Elo
44%W
·
50%L
·
6%T
1146
Elo
36%W
·
45%L
·
19%T
1128
Elo
38%W
·
50%L
·
12%T
Grok Imagine Image Pro
3 attempts · best result used for ranking
1194
Elo
56%W
·
33%L
·
11%T
1181
Elo
63%W
·
38%L
·
-1%T
1158
Elo
50%W
·
38%L
·
12%T
FLUX.2 [max]
3 attempts · best result used for ranking
1192
Elo
47%W
·
47%L
·
6%T
1142
Elo
43%W
·
43%L
·
14%T
1038
Elo
17%W
·
67%L
·
16%T
Nano Banana
3 attempts · best result used for ranking
1181
Elo
54%W
·
46%L
·
0%T
1164
Elo
33%W
·
67%L
·
0%T
1087
Elo
25%W
·
75%L
·
0%T
Reve Image 1.0
3 attempts · best result used for ranking
1138
Elo
33%W
·
61%L
·
6%T
1030
Elo
13%W
·
88%L
·
-1%T
995
Elo
0%W
·
100%L
·
0%T
Wan 2.6
3 attempts · best result used for ranking
1133
Elo
50%W
·
33%L
·
17%T
1127
Elo
31%W
·
62%L
·
7%T
1119
Elo
33%W
·
67%L
·
0%T
Original
Original 100%
Upscaled 100%
×
Hover the original to compare crops at 100%
Drag on the original to compare crops at 100%
RESULT
The most competitive head-to-head matchups, selected by closeness and vote count.
Google
OpenAI
ByteDance
Black Forest Labs
xAI
Reve AI
Alibaba
