Benchmark Almanac

SuperCLUE Arena Text to Image

ArtArk 1,494 tops its leaderboard GPT-Image-2.5 Flare 1,458 leads the models we track
Not in model cards

Models generate images from text. The arena ranks systems from pairwise comparisons.

The arena reports median ratings and uncertainty intervals.

Made by
SuperCLUE
Kind
Image generation and editing
Level
University
Scored
Ranked by matches
Answers
Open-ended
Tasks
All public
Measures
Preference
In and out
Text → Images
Interaction
Single turn
Items from
Live play or votes
Refresh
Live
Run by
A host
Last model added
November 2023
Scores refreshed
November 20th, 2026
28
Entries on its own leaderboard, 17 models we track

ArtArk tops the leaderboard with 1,494 GPT-Image-2.5 Flare leads with 1,458 among models released by November 2023

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is GPT-Image-2.5 Flare, at 1,458. This view shows only models we track, each at its best.

The newest model with a score here came out in November 2023, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

Bars start at 600.

  1. 1 ArtArk 1,494
  2. 2 GPT-Image-2.5 Flare 1,458
  3. 3 GPT-Image-2.5 Sunburst 1,441
  4. 4 GPT-Image-2 1,353
  5. 5 Gemini-3.1-Flash-Image-Preview(Nano Banana 2) 1,269
  6. 6 Seedream 5.0 Pro 1,235
  7. 7 Qwen-Image-3.0-Pro 1,172
  8. 8 Gemini 3 Pro Image Preview (nano-banana-pro) 1,164
  9. 9 GPT-Image-1.5 1,111
  10. 10 Seedream 4.5 1,049
  11. 11 Qwen-Image-2.0-Pro(2026-03-03) 1,032
  12. 12 Seedream 4.0 981
  13. 13 HunyuanImage 3.0 970
  14. 14 Seedream 3.0 936
  15. 15 SenseNova U1.5 Lite 933
  16. 16 Wan 2.6(2026-1-1) 925
  17. 17 Qwen-Image 898
  18. 18 Z-Image-Turbo 892
  19. 19 Wan 2.6(2025-12-16) 859
  20. 20 Seedream 5.0 Lite 855
  21. 21 GPT-Image-1 851
  22. 22 Claude 2.1 845
  23. 23 Imagen-4.0-Ultra 842
  24. 24 ERNIE-iRAG-1.0 783
  25. 25 wan2.2-t2i-plus 702
  26. 26 GLM-Image 663
  27. 27 wanx2.1-t2i-plus 651
  28. 28 CogView-4-250304 637

Source: Official SuperCLUE Arena Text to Image leaderboard

#EntryScore ±SettingDateSource
1 ArtArk 1,494 1447.8 to 1542.6 Official leaderboard
2 GPT-Image-2.5 Flare 1,458 1404.5 to 1508.1 Official leaderboard
3 GPT-Image-2.5 Sunburst 1,441 ±52.5 Official leaderboard
4 GPT-Image-2 1,353 1332.8 to 1376.1 Official leaderboard
5 Gemini-3.1-Flash-Image-Preview(Nano Banana 2) 1,269 ±18.1 Official leaderboard
6 Seedream 5.0 Pro 1,235 ±24 Official leaderboard
7 Qwen-Image-3.0-Pro 1,172 ±36.4 Official leaderboard
8 Gemini 3 Pro Image Preview (nano-banana-pro) 1,164 ±23.8 Official leaderboard
9 GPT-Image-1.5 1,111 1092.1 to 1130.8 Official leaderboard
10 Seedream 4.5 1,049 1024.2 to 1075.7 Official leaderboard
11 Qwen-Image-2.0-Pro(2026-03-03) 1,032 ±17.9 Official leaderboard
12 Seedream 4.0 981 958.6 to 1003.9 Official leaderboard
13 HunyuanImage 3.0 970 947.1 to 991.3 Official leaderboard
14 Seedream 3.0 936 911.4 to 957.1 Official leaderboard
15 SenseNova U1.5 Lite 933 897.8 to 969.2 Official leaderboard
16 Wan 2.6(2026-1-1) 925 ±20.9 Official leaderboard
17 Qwen-Image 898 874.2 to 921.2 Official leaderboard
18 Z-Image-Turbo 892 ±24.8 Official leaderboard
19 Wan 2.6(2025-12-16) 859 840.7 to 877.3 Official leaderboard
20 Seedream 5.0 Lite 855 838.6 to 871.3 Official leaderboard
21 GPT-Image-1 851 826.9 to 873.1 Official leaderboard
22 Claude 2.1 845 824.1 to 868.6 Nov 2023 Official leaderboard
23 Imagen-4.0-Ultra 842 ±22.6 Official leaderboard
24 ERNIE-iRAG-1.0 783 761.1 to 808 Official leaderboard
25 wan2.2-t2i-plus 702 ±22.6 Official leaderboard
26 GLM-Image 663 642.5 to 684.2 Official leaderboard
27 wanx2.1-t2i-plus 651 628.9 to 675.5 Official leaderboard
28 CogView-4-250304 637 613.1 to 661.7 Official leaderboard
#ModelScore ±SettingDateSource
1 GPT-Image-2.5 Flare 1,458 1404.5 to 1508.1 Official leaderboard
2 GPT-Image-2.5 Sunburst 1,441 ±52.5 Official leaderboard
3 GPT-Image-2 1,353 1332.8 to 1376.1 Official leaderboard
4 Seedream 5.0 Pro 1,235 ±24 Official leaderboard
5 Qwen-Image-3.0-Pro 1,172 ±36.4 Official leaderboard
6 Gemini 3 Pro Image Preview (nano-banana-pro) 1,164 ±23.8 Official leaderboard
7 GPT-Image-1.5 1,111 1092.1 to 1130.8 Official leaderboard
8 Seedream 4.5 1,049 1024.2 to 1075.7 Official leaderboard
9 Seedream 4.0 981 958.6 to 1003.9 Official leaderboard
10 HunyuanImage 3.0 970 947.1 to 991.3 Official leaderboard
11 Seedream 3.0 936 911.4 to 957.1 Official leaderboard
12 Qwen-Image 898 874.2 to 921.2 Official leaderboard
13 Z-Image-Turbo 892 ±24.8 Official leaderboard
14 Seedream 5.0 Lite 855 838.6 to 871.3 Official leaderboard
15 GPT-Image-1 851 826.9 to 873.1 Official leaderboard
16 Claude 2.1 845 824.1 to 868.6 Nov 2023 Official leaderboard
17 GLM-Image 663 642.5 to 684.2 Official leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard
845 2024 845 2024

Source: Official SuperCLUE Arena Text to Image leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Vision

All 636 in Vision
  • MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy. Qwen 3.6 Plus 86%
  • MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. GPT-5.6 Sol 84.6%
  • CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
  • MathVista In use Math reasoning in visual contexts (charts, geometry, puzzles) from 31 datasets. Answer accuracy. Seed2.1 Pro 90.7%
  • MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy. Qwen 3.8 Max 97.7%
  • MMStar In use Models answer 1,500 human-curated image questions that cannot be solved without the image, across 6 core capabilities. Scored by accuracy. Qwen 3.8 Max 85.9%

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 3 years ago, in November 2023, for Claude 2.1. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (28 entries).
  • Pass Models scored 17 models we track, plus 11 other entries on its leaderboard.
  • Warning Newest score The newest came 3 years ago, in November 2023, for Claude 2.1.
  • Note Frontier models None of its models are on the capabilities index.
  • Note Headroom Scored as a rating, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on superclueai.com: 17 rows for models we track and 11 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check