Text-to-Image ArenaactiveImage generation

Text-to-Image Arena: LMArena Text-to-Image Arena

Metric: Arena Score (higher is better)Introduced: 2025

Crowdsourced human-preference leaderboard for text-to-image models run by LMArena. Voters compare two anonymous images generated from the same prompt and pick the better one; ratings use a Bradley-Terry model with 95% confidence intervals. Higher is better.

Configurations

34

tracked on this benchmark

Score band

1425.0 – 938.0

best → lowest tracked

Snapshot trend

—

need ≥2 snapshots

Leaderboard

Tracked models ranked by Arena Score (higher is better).

Compare candidates
#Published configuration and provenanceScore
1

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Qualification: Preliminary (LMArena)

Notes: Arena score 1425 (95% CI 1418-1431), 17,752 votes, rank 1 of 82, category overall, leaderboard published 2026-10-07. Board entry 'gpt-image-2.5-sunburst'. LMArena marks this score Preliminary. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,425
2

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Qualification: Preliminary (LMArena)

Notes: Arena score 1398 (95% CI 1391-1404), 16,889 votes, rank 2 of 82, category overall, leaderboard published 2026-10-07. Board entry 'gpt-image-2.5-flare'. LMArena marks this score Preliminary. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,398
3

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Qualification: Preliminary (LMArena)

Notes: Arena score 1328 (95% CI 1319-1337), 5,310 votes, rank 6 of 82, category overall, leaderboard published 2026-10-07. Board entry 'gemini-nano-banana-2.1'. LMArena marks this score Preliminary. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,328
4

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1302 (95% CI 1294-1309), 8,249 votes, rank 7 of 82, category overall, leaderboard published 2026-10-07. Board entry 'reve-2.1'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,302
5

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1269 (95% CI 1263-1275), 15,773 votes, rank 10 of 82, category overall, leaderboard published 2026-10-07. Board entry 'reve-2.0'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,269
6

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1255 (95% CI 1249-1262), 11,440 votes, rank 13 of 82, category overall, leaderboard published 2026-10-07. Board entry 'qwen-image-3.0-pro'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,255
7

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1254 (95% CI 1250-1258), 61,265 votes, rank 14 of 82, category overall, leaderboard published 2026-10-07. Board entry 'mai-image-2.5'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,254
8

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1234 (95% CI 1229-1239), 86,464 votes, rank 18 of 82, category overall, leaderboard published 2026-10-07. Board entry 'gemini-3-pro-image-preview (nano-banana-pro)'. LLM Reference model is lifecycle-deprecated; score kept as historical. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,234
9
Qwen-Image-2.1Preliminary

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Qualification: Preliminary (LMArena)

Notes: Arena score 1223 (95% CI 1216-1230), 11,441 votes, rank 19 of 82, category overall, leaderboard published 2026-10-07. Board entry 'qwen-image-2.1'. LMArena marks this score Preliminary. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,223
10

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1184 (95% CI 1179-1188), 51,475 votes, rank 24 of 82, category overall, leaderboard published 2026-10-07. Board entry 'mai-image-2'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,184
11

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1170 (95% CI 1167-1172), 272,085 votes, rank 25 of 82, category overall, leaderboard published 2026-10-07. Board entry 'grok-imagine-image'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,170
12

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1162 (95% CI 1159-1166), 99,455 votes, rank 28 of 82, category overall, leaderboard published 2026-10-07. Board entry 'grok-imagine-image-pro'. LLM Reference model is lifecycle-deprecated; score kept as historical. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,162
13

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1152 (95% CI 1146-1157), 18,485 votes, rank 33 of 82, category overall, leaderboard published 2026-10-07. Board entry 'Cosmos3-Super-Text2Image'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,152
14

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1150 (95% CI 1146-1153), 180,518 votes, rank 35 of 82, category overall, leaderboard published 2026-10-07. Board entry 'hunyuan-image-3.0'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,150
15

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1149 (95% CI 1146-1151), 300,107 votes, rank 36 of 82, category overall, leaderboard published 2026-10-07. Board entry 'seedream-4.5'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,149
16

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1148 (95% CI 1143-1152), 388,736 votes, rank 37 of 82, category overall, leaderboard published 2026-10-07. Board entry 'imagen-ultra-4.0-generate-001'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,148
17

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1137 (95% CI 1135-1140), 233,292 votes, rank 41 of 82, category overall, leaderboard published 2026-10-07. Board entry 'wan2.6-t2i'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,137
18

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1129 (95% CI 1126-1132), 539,732 votes, rank 43 of 82, category overall, leaderboard published 2026-10-07. Board entry 'imagen-4.0-generate-001'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,129
19

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1116 (95% CI 1112-1119), 268,453 votes, rank 49 of 82, category overall, leaderboard published 2026-10-07. Board entry 'gpt-image-1'. LLM Reference model is lifecycle-deprecated; score kept as historical. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,116
20

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1115 (95% CI 1112-1118), 137,017 votes, rank 50 of 82, category overall, leaderboard published 2026-10-07. Board entry 'recraft-v4'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,115
21

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1110 (95% CI 1107-1113), 171,058 votes, rank 54 of 82, category overall, leaderboard published 2026-10-07. Board entry 'gpt-image-1-mini'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,110
22

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1059 (95% CI 1055-1062), 332,367 votes, rank 64 of 82, category overall, leaderboard published 2026-10-07. Board entry 'flux-1-kontext-pro'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,059
23

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1058 (95% CI 1055-1061), 359,772 votes, rank 65 of 82, category overall, leaderboard published 2026-10-07. Board entry 'imagen-3.0-generate-002'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,058
24

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1057 (95% CI 1054-1059), 84,550 votes, rank 66 of 82, category overall, leaderboard published 2026-10-07. Board entry 'qwen-image'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,057
25

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1036 (95% CI 1032-1039), 130,160 votes, rank 68 of 82, category overall, leaderboard published 2026-10-07. Board entry 'photon'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,036
26

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1024 (95% CI 1020-1029), 54,361 votes, rank 71 of 82, category overall, leaderboard published 2026-10-07. Board entry 'runway-gen4'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,024
27

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1021 (95% CI 1017-1024), 195,627 votes, rank 72 of 82, category overall, leaderboard published 2026-10-07. Board entry 'recraft-v3'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,021
28

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1016 (95% CI 1012-1019), 70,460 votes, rank 73 of 82, category overall, leaderboard published 2026-10-07. Board entry 'flux-1.1-pro'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,016
29

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1013 (95% CI 1010-1017), 72,090 votes, rank 74 of 82, category overall, leaderboard published 2026-10-07. Board entry 'ideogram-v2'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,013
30

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1013 (95% CI 1010-1016), 287,341 votes, rank 75 of 82, category overall, leaderboard published 2026-10-07. Board entry 'lucid-origin'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,013
31

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 1012 (95% CI 1003-1021), 4,854 votes, rank 76 of 82, category overall, leaderboard published 2026-10-07. Board entry 'glm-image'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

1,012
32

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 975 (95% CI 972-979), 257,227 votes, rank 77 of 82, category overall, leaderboard published 2026-10-07. Board entry 'gemini-2.0-flash-preview-image-generation'. LLM Reference model is lifecycle-deprecated; score kept as historical. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

975
33

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 968 (95% CI 964-972), 238,743 votes, rank 79 of 82, category overall, leaderboard published 2026-10-07. Board entry 'dall-e-3'. LLM Reference model is lifecycle-deprecated; score kept as historical. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

968
34

Configuration: Overall leaderboard entry (unqualified board entry)

Version: text_to_image overall 2026-10-07Harness: LMArena blind pairwise human votes, Bradley-TerryEvaluator: LMArenaObserved: Oct 7, 2026Confidence: confirmedSource

Notes: Arena score 938 (95% CI 934-943), 23,396 votes, rank 81 of 82, category overall, leaderboard published 2026-10-07. Board entry 'stable-diffusion-v35-large'. Source: LMArena (CC BY 4.0). Retrieved 2026-10-17.

938

Attribution. Source: LMArena (lmarena.ai), CC BY 4.0

Scores rounded; model entries mapped to LLM Reference models; snapshot 2026-10-07.

How to read this benchmark

This benchmark scores models where higher is better. Use scores for directional filtering and shortlisting, not universal quality ranking; then validate pricing, context window, provider availability, and fit for your workload.

Trust this score when

  • The row names the exact model configuration, benchmark version, evaluator, harness, and observation date.
  • Cost, token, and step averages are evaluated together with the score for the same configuration.
  • You validate the configuration in your own Cursor workload before making a model choice.

Be cautious when

  • Rows are collapsed into one score per base model or compared across benchmark versions.
  • Small score differences are treated as decisive despite Cursor's variance caveat.
  • A qualified row, including Grok 4.5's disclosed training-data advantage, is used as neutral winner evidence.

Related benchmarks

Last reviewed: Oct 17, 2026