Valumigo

LMArena · Text in images

We regularly fetch and display public benchmark data: LMArena user voting rankings, Epoch AI test scores, and OpenRouter usage rankings. Each leaderboard shows both its publication date and the date we retrieved it. These scores are not assigned by this site.

LMArena scores come from people comparing two models' answers side by side and voting for the better one, using an Elo-based system. If score differences fall within the confidence intervals, treat the models as roughly comparable.

How to read this leaderboard

What it measures: A ranking based on preference votes for prompts explicitly requesting readable, accurate text within images.

How to read the score: A relative score reflecting which images users preferred for these prompts. It is not an accuracy rate that separately scores text correctness.

Caveats: This ranking alone cannot establish performance in rendering Korean (Hangul) characters. Check separately using Korean (Hangul) prompts.

Insights from this ranking

  • The 95% confidence intervals for 1st-place gpt-image-2.5-sunburst and 2nd-place gpt-image-2.5-flare do not overlap. This aggregation provides relatively clear evidence that the 1st-place model leads.
  • OpenAI has the most models among the top 10, with 3.
  • The 1st-place model received 3,699 votes, and this leaderboard includes 50 models.

Text in images LMArena

Published 2026-09-24 · Retrieved 2026-10-04

112012001280136014401520
Ogpt-image-2.5-sunburst
1467
Ogpt-image-2.5-flare
1435
Ogpt-image-2 (medium)
1425
Mmai-image-2.6
1371
xgrok-imagine-image-2.0 (low)
1341
Rreve-2.1
1340
Metamuse-image
1299
Rreve-2.0
1299
Googlegemini-3.1-flash-image (nano-banana-2) [web-search]
1295
Alibabaqwen-image-3.0-pro
1283
Mmai-image-2.5
1281
Googlegemini-3.1-flash-lite-image (nano-banana-2-lite)
1274
Googlegemini-3-pro-image-2k (nano-banana-pro)
1271
Alibabaqwen-image-2.1
1262
ByteDanceseedream-5.0-pro
1255
Googlegemini-3-pro-image-preview (nano-banana-pro)
1254
Ogpt-image-1.5-high-fidelity
1253
Iideogram-4.0-quality
1238
Luni-1.1-max
1219
Alibabaqwen-image-2.0-pro-2026-06-22
1202

Dots show scores; horizontal lines show 95% confidence intervals. Overlapping intervals suggest similar performance. The chart shows only the highest-scoring reasoning setting, such as high or max, for each model. The table shows rankings for all settings.

Top ranking by company

  • OOpenAIgpt-image-2.5-sunburst#1
  • MMicrosoft AImai-image-2.6#4
  • xxAIgrok-imagine-image-2.0 (low)#5
  • RRevereve-2.1#6
  • MetaMetamuse-image#7
  • GoogleGooglegemini-3.1-flash-image (nano-banana-2) [web-search]#9
  • AlibabaAlibabaqwen-image-3.0-pro#10
  • ByteDanceByteDanceseedream-5.0-pro#15
  • IIdeogramideogram-4.0-quality#18
View table (top 50)
RankModelDeveloperScore95% confidence intervalVotes
1#1Ogpt-image-2.5-sunburstOpenAI14671455–14803,699
2#2Ogpt-image-2.5-flareOpenAI14351423–14483,360
3#3Ogpt-image-2 (medium)OpenAI14251419–143232,795
4#4Mmai-image-2.6Microsoft AI13711362–13816,053
5#5xgrok-imagine-image-2.0 (low)xAI13411329–13542,354
6#6Rreve-2.1Reve13401328–13522,756
7#7Metamuse-imageMeta12991292–130711,345
8#8Rreve-2.0Reve12991290–13085,468
9#9Googlegemini-3.1-flash-image (nano-banana-2) [web-search]Google12951289–130117,342
10#10Alibabaqwen-image-3.0-proAlibaba12831273–12933,643
11#11Mmai-image-2.5Microsoft AI12811275–128720,912
12#12Googlegemini-3.1-flash-lite-image (nano-banana-2-lite)Google12741265–12835,826
13#13Googlegemini-3-pro-image-2k (nano-banana-pro)Google12711266–127555,545
14#14Alibabaqwen-image-2.1Alibaba12621247–12781,361
15#15ByteDanceseedream-5.0-proByteDance12551249–126127,232
16#16Googlegemini-3-pro-image-preview (nano-banana-pro)Google12541247–126225,697
17#17Ogpt-image-1.5-high-fidelityOpenAI12531248–125859,309
18#18Iideogram-4.0-qualityIdeogram12381232–124417,299
19#19Luni-1.1-maxLuma AI12191210–12284,796
20#20Luni-1.1Luma AI12071201–121413,698
21#21Alibabaqwen-image-2.0-pro-2026-06-22Alibaba12021193–12114,374
22#22Rreve-v1.5Reve11871181–119313,637
23#23xgrok-imagine-imagexAI11841180–118993,401
24#24xgrok-imagine-image-proxAI11821177–118835,037
25#25Mmai-image-2Microsoft AI11771170–118319,396
26#26Bflux-2-flexBlack Forest Labs11721167–117752,342
27#27Bflux-2-maxBlack Forest Labs11661160–117142,386
28#28Rrecraft-v4.1-utility-proRecraft11621144–1179956
29#29Bflux-2-devBlack Forest Labs11581153–116424,410
30#30Rrecraft-v4.1-proRecraft11561139–1173956
31#31Bflux-2-proBlack Forest Labs11551151–116069,022
32#32Googlegemini-2.5-flash-image-preview (nano-banana)Google11541150–1157234,500
33#33Thunyuan-image-3.0Tencent11491144–115452,522
34#34Alibabawan2.6-t2iAlibaba11491144–115379,115
35#35Googleimagen-ultra-4.0-generate-001Google11481143–115490,345
36#36NVIDIACosmos3-Super-Text2Image (Agentic)NVIDIA11451134–11572,436
37#37ByteDanceseedream-4.5ByteDance11431139–114797,172
38#38ByteDanceseedream-4-2kByteDance11431131–11543,099
39#39Alibabawan2.5-t2i-previewAlibaba11421138–114688,333
40#40Rrecraft-v4Recraft11391134–114445,141
41#41ByteDanceseedream-5.0-liteByteDance11351131–114043,258
42#42Hhidream-o1-imageHidream11341129–114018,944
43#43Kkrea-2-mediumKrea11271121–113416,666
44#44ByteDanceseedream-4-falByteDance11241113–11352,747
45#45Alibabaqwen-image-2512Alibaba11231118–112830,851
46#46Googleimagen-4.0-generate-001Google11231119–1127141,558
47#47Alibabawan2.7-image-proAlibaba11211114–112712,680
48#48NVIDIACosmos3-Super-Text2ImageNVIDIA11191110–11285,365
49#49Alibabawan2.7-imageAlibaba11161109–112312,705
50#50Ogpt-image-1OpenAI11151110–111972,150

Data: LMArena leaderboard dataset (CC BY 4.0). Changes: Top 50 entries per leaderboard; scores are rounded. https://huggingface.co/datasets/lmarena-ai/leaderboard-datasetCompany logos are trademarks of their respective owners and are used only for identification (icons: Simple Icons).Data: Epoch AI — Capabilities & benchmarking (CC BY 4.0). https://epoch.ai/benchmarksSource: OpenRouter (openrouter.ai/rankings), as of 2026-10-04. Licensed under CC BY 4.0. https://openrouter.ai/rankings