Benchmark Almanac

LMArena Text-to-Image Arena

GPT-Image-2.5 Sunburst 1,429 leads since November 2026
Not in model cards

Image generation models answer the same user prompt side by side and anonymously. Community votes between the two images give Elo-style ratings.

LMArena announced the arena in December 2024 and opened the leaderboard in January 2025, per posts found in search. Ratings are in the text_to_image config of the dataset.

Released
December 2024
Kind
Image generation and editing
Level
Everyday
Scored
Ranked by votes
Answers
Open-ended
Tasks
All public
Last model added
November 2026
Scores refreshed
November 12th, 2026
79
Entries on its own leaderboard, 78 models we track

GPT-Image-2.5 Sunburst leads with 1,429

Best score per entry from the benchmark's own leaderboard, read on November 12th, 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.

79 entries. Scroll sideways for the rest. Bars start at 800.

  1. 1,429 GPT-Image-2.5 Sunburst
  2. 1,406 GPT-Image-2.6 Sunburst
  3. 1,400 GPT-Image-2.5 Flare
  4. 1,383 GPT-Image-2.6 Flare
  5. 1,382 GPT-Image-2
  6. 1,329 MAI-Image-2.6
  7. 1,328 Gemini Nano Banana 2.1
  8. 1,319 Grok Imagine Image 2.0
  9. 1,302 Reve 2.1
  10. 1,268 Reve 2.0
  11. 1,266 Muse Image
  12. 1,257 Gemini 3.1 Flash Image (nano-banana-2, Web Search)
  13. 1,254 Qwen-Image-3.0-Pro
  14. 1,253 MAI-Image-2.5
  15. 1,249 Gemini 3.1 Flash-Lite Image (nano-banana-2-lite)
  16. 1,249 Seedream 5.0 Pro
  17. 1,246 Gemini 3 Pro Image (nano-banana-pro, 2K)
  18. 1,236 GPT-Image-1.5
  19. 1,231 Gemini 3 Pro Image Preview (nano-banana-pro)
  20. 1,227 Qwen-Image-2.1
  21. 1,199 Ideogram 4.0
  22. 1,190 Qwen-Image-2.0-Pro
  23. 1,188 UNI-1.1 Max
  24. 1,183 MAI-Image-2
  25. 1,182 UNI-1.1
  26. 1,169 Recraft V4.1 Utility Pro
  27. 1,168 Grok Imagine Image
  28. 1,164 Cosmos3-Super-Text2Image
  29. 1,161 FLUX.2 [max]
  30. 1,160 Grok Imagine Image Pro
  31. 1,156 FLUX.2 [flex]
  32. 1,154 Reve v1.5
  33. 1,153 FLUX.2 [pro]
  34. 1,152 HunyuanImage 3.0
  35. 1,149 Gemini 2.5 Flash Image Preview (nano-banana)
  36. 1,148 Seedream 4.5
  37. 1,146 Imagen 4 Ultra (generate-001)
  38. 1,145 FLUX.2 [dev]
  39. 1,141 Seedream 5.0 Lite
  40. 1,138 Wan 2.6
  41. 1,138 Seedream 4.0
  42. 1,132 Recraft V4.1 Pro
  43. 1,128 Krea 2 Medium
  44. 1,128 Imagen 4 (generate-001)
  45. 1,124 Qwen-Image (2512)
  46. 1,117 Wan 2.5 Preview
  47. 1,116 Recraft V4
  48. 1,116 Krea 2 Turbo
  49. 1,115 GPT-Image-1
  50. 1,115 HiDream-O1-Image
  51. 1,114 Krea 2 Large
  52. 1,109 GPT-Image-1 mini
  53. 1,103 Wan 2.7 Image Pro
  54. 1,100 Wan 2.7 Image
  55. 1,098 Recraft V4.1 Flash
  56. 1,093 MAI-Image-1
  57. 1,082 Z-Image-Turbo
  58. 1,081 Seedream 3.0
  59. 1,075 FLUX.1 Kontext [max]
  60. 1,070 FLUX.2 [klein] 9B
  61. 1,061 Qwen-Image
  62. 1,059 FLUX.1 Kontext [pro]
  63. 1,057 Imagen 3 (generate-002)
  64. 1,049 Ideogram 3.0
  65. 1,036 Luma Photon
  66. 1,033 P-Image
  67. 1,030 FLUX.2 [klein] 4B
  68. 1,024 Runway Gen-4
  69. 1,022 Recraft V3
  70. 1,016 FLUX1.1 [pro]
  71. 1,013 Ideogram 2.0
  72. 1,013 Lucid Origin
  73. 1,012 GLM-Image
  74. 976 Gemini 2.0 Flash Preview Image Generation
  75. 969 FLUX.1 [dev]
  76. 968 DALL-E 3
  77. 941 FLUX.1 Kontext [dev]
  78. 938 Stable Diffusion 3.5 Large
  79. 897 BAGEL

Bars start at 800.

  1. 1 GPT-Image-2.5 Sunburst 1,429
  2. 2 GPT-Image-2.6 Sunburst medium 1,406
  3. 3 GPT-Image-2.5 Flare 1,400
  4. 4 GPT-Image-2.6 Flare medium 1,383
  5. 5 GPT-Image-2 medium 1,382
  6. 6 MAI-Image-2.6 1,329
  7. 7 Gemini Nano Banana 2.1 1,328
  8. 8 Grok Imagine Image 2.0 1,319
  9. 9 Reve 2.1 1,302
  10. 10 Reve 2.0 1,268
  11. 11 Muse Image 1,266
  12. 12 Gemini 3.1 Flash Image (nano-banana-2, Web Search) 1,257
  13. 13 Qwen-Image-3.0-Pro 1,254
  14. 14 MAI-Image-2.5 1,253
  15. 15 Gemini 3.1 Flash-Lite Image (nano-banana-2-lite) 1,249
  16. 16 Seedream 5.0 Pro 1,249
  17. 17 Gemini 3 Pro Image (nano-banana-pro, 2K) 1,246
  18. 18 GPT-Image-1.5 high fidelity 1,236
  19. 19 Gemini 3 Pro Image Preview (nano-banana-pro) 1,231
  20. 20 Qwen-Image-2.1 1,227
  21. 21 Ideogram 4.0 1,199
  22. 22 Qwen-Image-2.0-Pro 1,190
  23. 23 UNI-1.1 Max 1,188
  24. 24 MAI-Image-2 1,183
  25. 25 UNI-1.1 1,182
  26. 26 Recraft V4.1 Utility Pro 1,169
  27. 27 Grok Imagine Image 1,168
  28. 28 Cosmos3-Super-Text2Image 1,164
  29. 29 FLUX.2 [max] 1,161
  30. 30 Grok Imagine Image Pro 1,160
  31. 31 FLUX.2 [flex] 1,156
  32. 32 Reve v1.5 1,154
  33. 33 FLUX.2 [pro] 1,153
  34. 34 HunyuanImage 3.0 1,152
  35. 35 Gemini 2.5 Flash Image Preview (nano-banana) 1,149
  36. 36 Seedream 4.5 1,148
  37. 37 Imagen 4 Ultra (generate-001) 1,146
  38. 38 FLUX.2 [dev] 1,145
  39. 39 Seedream 5.0 Lite 1,141
  40. 40 Wan 2.6 1,138
  41. 41 Seedream 4.0 1,138
  42. 42 Recraft V4.1 Pro 1,132
  43. 43 Krea 2 Medium 1,128
  44. 44 Imagen 4 (generate-001) 1,128
  45. 45 Qwen-Image (2512) 1,124
  46. 46 Wan 2.5 Preview 1,117
  47. 47 Recraft V4 1,116
  48. 48 Krea 2 Turbo 1,116
  49. 49 GPT-Image-1 1,115
  50. 50 HiDream-O1-Image 1,115
  51. 51 Krea 2 Large 1,114
  52. 52 GPT-Image-1 mini 1,109
  53. 53 Wan 2.7 Image Pro 1,103
  54. 54 Wan 2.7 Image 1,100
  55. 55 Recraft V4.1 Flash 1,098
  56. 56 MAI-Image-1 1,093
  57. 57 Z-Image-Turbo 1,082
  58. 58 Seedream 3.0 1,081
  59. 59 FLUX.1 Kontext [max] 1,075
  60. 60 FLUX.2 [klein] 9B 1,070
  61. 61 Qwen-Image 1,061
  62. 62 FLUX.1 Kontext [pro] 1,059
  63. 63 Imagen 3 (generate-002) 1,057
  64. 64 Ideogram 3.0 1,049
  65. 65 Luma Photon 1,036
  66. 66 P-Image 1,033
  67. 67 FLUX.2 [klein] 4B 1,030
  68. 68 Runway Gen-4 1,024
  69. 69 Recraft V3 1,022
  70. 70 FLUX1.1 [pro] 1,016
  71. 71 Ideogram 2.0 1,013
  72. 72 Lucid Origin 1,013
  73. 73 GLM-Image 1,012
  74. 74 Gemini 2.0 Flash Preview Image Generation 976
  75. 75 FLUX.1 [dev] fp8 969
  76. 76 DALL-E 3 968
  77. 77 FLUX.1 Kontext [dev] 941
  78. 78 Stable Diffusion 3.5 Large 938
  79. 79 BAGEL 897

Source: Official LMArena Text-to-Image Arena leaderboard

#EntryScoreSettingDateSource
1 GPT-Image-2.5 Sunburst 1,429 Nov 2026 Official leaderboard
2 GPT-Image-2.6 Sunburst 1,406 medium Nov 2026 Official leaderboard
3 GPT-Image-2.5 Flare 1,400 Nov 2026 Official leaderboard
4 GPT-Image-2.6 Flare 1,383 medium Nov 2026 Official leaderboard
5 GPT-Image-2 1,382 medium Nov 2026 Official leaderboard
6 MAI-Image-2.6 1,329 Nov 2026 Official leaderboard
7 Gemini Nano Banana 2.1 1,328 Nov 2026 Official leaderboard
8 Grok Imagine Image 2.0 1,319 Nov 2026 Official leaderboard
9 Reve 2.1 1,302 Nov 2026 Official leaderboard
10 Reve 2.0 1,268 Nov 2026 Official leaderboard
11 Muse Image 1,266 Nov 2026 Official leaderboard
12 Gemini 3.1 Flash Image (nano-banana-2, Web Search) 1,257 Nov 2026 Official leaderboard
13 Qwen-Image-3.0-Pro 1,254 Nov 2026 Official leaderboard
14 MAI-Image-2.5 1,253 Nov 2026 Official leaderboard
15 Gemini 3.1 Flash-Lite Image (nano-banana-2-lite) 1,249 Nov 2026 Official leaderboard
16 Seedream 5.0 Pro 1,249 Nov 2026 Official leaderboard
17 Gemini 3 Pro Image (nano-banana-pro, 2K) 1,246 Nov 2026 Official leaderboard
18 GPT-Image-1.5 1,236 high fidelity Nov 2026 Official leaderboard
19 Gemini 3 Pro Image Preview (nano-banana-pro) 1,231 Nov 2026 Official leaderboard
20 Qwen-Image-2.1 1,227 Nov 2026 Official leaderboard
21 Ideogram 4.0 1,199 Nov 2026 Official leaderboard
22 Qwen-Image-2.0-Pro 1,190 Jun 2026 Official leaderboard
23 UNI-1.1 Max 1,188 Nov 2026 Official leaderboard
24 MAI-Image-2 1,183 Nov 2026 Official leaderboard
25 UNI-1.1 1,182 Nov 2026 Official leaderboard
26 Recraft V4.1 Utility Pro 1,169 Nov 2026 Official leaderboard
27 Grok Imagine Image 1,168 Nov 2026 Official leaderboard
28 Cosmos3-Super-Text2Image 1,164 Nov 2026 Official leaderboard
29 FLUX.2 [max] 1,161 Nov 2026 Official leaderboard
30 Grok Imagine Image Pro 1,160 Nov 2026 Official leaderboard
31 FLUX.2 [flex] 1,156 Nov 2026 Official leaderboard
32 Reve v1.5 1,154 Nov 2026 Official leaderboard
33 FLUX.2 [pro] 1,153 Nov 2026 Official leaderboard
34 HunyuanImage 3.0 1,152 Nov 2026 Official leaderboard
35 Gemini 2.5 Flash Image Preview (nano-banana) 1,149 Nov 2026 Official leaderboard
36 Seedream 4.5 1,148 Nov 2026 Official leaderboard
37 Imagen 4 Ultra (generate-001) 1,146 Nov 2026 Official leaderboard
38 FLUX.2 [dev] 1,145 Nov 2026 Official leaderboard
39 Seedream 5.0 Lite 1,141 Nov 2026 Official leaderboard
40 Wan 2.6 1,138 Nov 2026 Official leaderboard
41 Seedream 4.0 1,138 Nov 2026 Official leaderboard
42 Recraft V4.1 Pro 1,132 Nov 2026 Official leaderboard
43 Krea 2 Medium 1,128 Nov 2026 Official leaderboard
44 Imagen 4 (generate-001) 1,128 Nov 2026 Official leaderboard
45 Qwen-Image (2512) 1,124 Nov 2026 Official leaderboard
46 Wan 2.5 Preview 1,117 Nov 2026 Official leaderboard
47 Recraft V4 1,116 Nov 2026 Official leaderboard
48 Krea 2 Turbo 1,116 Nov 2026 Official leaderboard
49 GPT-Image-1 1,115 Nov 2026 Official leaderboard
50 HiDream-O1-Image 1,115 Nov 2026 Official leaderboard
51 Krea 2 Large 1,114 Nov 2026 Official leaderboard
52 GPT-Image-1 mini 1,109 Nov 2026 Official leaderboard
53 Wan 2.7 Image Pro 1,103 Nov 2026 Official leaderboard
54 Wan 2.7 Image 1,100 Nov 2026 Official leaderboard
55 Recraft V4.1 Flash 1,098 Nov 2026 Official leaderboard
56 MAI-Image-1 1,093 Nov 2026 Official leaderboard
57 Z-Image-Turbo 1,082 Nov 2026 Official leaderboard
58 Seedream 3.0 1,081 Nov 2026 Official leaderboard
59 FLUX.1 Kontext [max] 1,075 Nov 2026 Official leaderboard
60 FLUX.2 [klein] 9B 1,070 Nov 2026 Official leaderboard
61 Qwen-Image 1,061 Nov 2026 Official leaderboard
62 FLUX.1 Kontext [pro] 1,059 Nov 2026 Official leaderboard
63 Imagen 3 (generate-002) 1,057 Nov 2026 Official leaderboard
64 Ideogram 3.0 1,049 Nov 2026 Official leaderboard
65 Luma Photon 1,036 Nov 2026 Official leaderboard
66 P-Image 1,033 Nov 2026 Official leaderboard
67 FLUX.2 [klein] 4B 1,030 Nov 2026 Official leaderboard
68 Runway Gen-4 1,024 Nov 2026 Official leaderboard
69 Recraft V3 1,022 Nov 2026 Official leaderboard
70 FLUX1.1 [pro] 1,016 Nov 2026 Official leaderboard
71 Ideogram 2.0 1,013 Nov 2026 Official leaderboard
72 Lucid Origin 1,013 Nov 2026 Official leaderboard
73 GLM-Image 1,012 Nov 2026 Official leaderboard
74 Gemini 2.0 Flash Preview Image Generation 976 Nov 2026 Official leaderboard
75 FLUX.1 [dev] 969 fp8 Nov 2026 Official leaderboard
76 DALL-E 3 968 Nov 2026 Official leaderboard
77 FLUX.1 Kontext [dev] 941 Nov 2026 Official leaderboard
78 Stable Diffusion 3.5 Large 938 Nov 2026 Official leaderboard
79 BAGEL 897 Nov 2026 Official leaderboard

4 models have held number one on the official leaderboard since its release in December 2024

Lab claims count from the day the card came out, independent runs from the model’s release.

Qwen xAI OpenAI
Own leaderboard 4 leaders
GPT-Image-2.5 Sunburst 1,429

Source: Official LMArena Text-to-Image Arena 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
800 1,000 1,200 1,400 1,600 2026 2027 800 1,000 1,200 1,400 1,600 2026 2027

Source: Official LMArena Text-to-Image Arena leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Vision

All 281 in Vision
  • MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy. GPT-5.1 85.4%
  • 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%
  • 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%
  • ChartQA In use Models answer human-written and generated questions about chart images that need visual and arithmetic reasoning. Scored by relaxed accuracy with a 5 percent numeric tolerance. Command A Vision 90.9%