LMArena Text-to-Image Arena
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.
- Level
- Everyday Most adults, without training
- Scored
- Ranked by votes People pick the better of two answers, as in an arena, and the votes are turned into ratings.
- Answers
- Open-ended Free-form text, documents, designs or media judged as a whole.
GPT-Image-2.5 Sunburst leads with 1,427
Best score per entry from the benchmark's own leaderboard, read on 2 Nov 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.
- 1,427 GPT-Image-2.5 Sunburst
- 1,401 GPT-Image-2.5 Flare
- 1,381 GPT-Image-2
- 1,329
MAI-Image-2.6
- 1,325
Gemini Nano Banana 2.1
- 1,317 Grok Imagine Image 2.0
- 1,300 Reve 2.1
- 1,269
Muse Image
- 1,268 Reve 2.0
- 1,257
Gemini 3.1 Flash Image (nano-banana-2, Web Search)
- 1,254
Qwen-Image-3.0-Pro
- 1,253
MAI-Image-2.5
- 1,250
Seedream 5.0 Pro
- 1,249
Gemini 3.1 Flash-Lite Image (nano-banana-2-lite)
- 1,246
Gemini 3 Pro Image (nano-banana-pro, 2K)
- 1,236 GPT-Image-1.5
- 1,232
Gemini 3 Pro Image Preview (nano-banana-pro)
- 1,225
Qwen-Image-2.1
- 1,200 Ideogram 4.0
- 1,191
Qwen-Image-2.0-Pro
- 1,189 UNI-1.1 Max
- 1,183
MAI-Image-2
- 1,183 UNI-1.1
- 1,169 Grok Imagine Image
- 1,168 Recraft V4.1 Utility Pro
- 1,164
Cosmos3-Super-Text2Image
- 1,161 FLUX.2 [max]
- 1,161 Grok Imagine Image Pro
- 1,156 FLUX.2 [flex]
- 1,154 Reve v1.5
- 1,153 FLUX.2 [pro]
- 1,151
HunyuanImage 3.0
- 1,149
Seedream 4.5
- 1,149
Gemini 2.5 Flash Image Preview (nano-banana)
- 1,146
Imagen 4 Ultra (generate-001)
- 1,145 FLUX.2 [dev]
- 1,140
Seedream 5.0 Lite
- 1,139
Seedream 4.0
- 1,138
Wan 2.6
- 1,131 Recraft V4.1 Pro
- 1,128
Imagen 4 (generate-001)
- 1,128 Krea 2 Medium
- 1,125
Qwen-Image (2512)
- 1,117
Wan 2.5 Preview
- 1,115 Krea 2 Large
- 1,115 GPT-Image-1
- 1,115 Recraft V4
- 1,115
HiDream-O1-Image
- 1,114 Krea 2 Turbo
- 1,109 GPT-Image-1 mini
- 1,103
Wan 2.7 Image Pro
- 1,101
Wan 2.7 Image
- 1,098 Recraft V4.1 Flash
- 1,092
MAI-Image-1
- 1,082
Z-Image-Turbo
- 1,082
Seedream 3.0
- 1,075 FLUX.1 Kontext [max]
- 1,070 FLUX.2 [klein] 9B
- 1,061
Qwen-Image
- 1,059 FLUX.1 Kontext [pro]
- 1,057
Imagen 3 (generate-002)
- 1,049 Ideogram 3.0
- 1,036 Luma Photon
- 1,032 P-Image
- 1,030 FLUX.2 [klein] 4B
- 1,024 Runway Gen-4
- 1,022 Recraft V3
- 1,016 FLUX1.1 [pro]
- 1,013 Ideogram 2.0
- 1,013
Lucid Origin
- 1,012 GLM-Image
- 976
Gemini 2.0 Flash Preview Image Generation
- 969 FLUX.1 [dev]
- 967 DALL-E 3
- 942 FLUX.1 Kontext [dev]
- 938
Stable Diffusion 3.5 Large
- 897
BAGEL
Bars start at 800.
- 1 GPT-Image-2.5 Sunburst 1,427
- 2 GPT-Image-2.5 Flare 1,401
- 3 GPT-Image-2 medium 1,381
- 4
MAI-Image-2.6 1,329
- 5
Gemini Nano Banana 2.1 1,325
- 6 Grok Imagine Image 2.0 1,317
- 7 Reve 2.1 1,300
- 8
Muse Image 1,269
- 9 Reve 2.0 1,268
- 10
Gemini 3.1 Flash Image (nano-banana-2, Web Search) 1,257
- 11
Qwen-Image-3.0-Pro 1,254
- 12
MAI-Image-2.5 1,253
- 13
Seedream 5.0 Pro 1,250
- 14
Gemini 3.1 Flash-Lite Image (nano-banana-2-lite) 1,249
- 15
Gemini 3 Pro Image (nano-banana-pro, 2K) 1,246
- 16 GPT-Image-1.5 high fidelity 1,236
- 17
Gemini 3 Pro Image Preview (nano-banana-pro) 1,232
- 18
Qwen-Image-2.1 1,225
- 19 Ideogram 4.0 1,200
- 20
Qwen-Image-2.0-Pro 1,191
- 21 UNI-1.1 Max 1,189
- 22
MAI-Image-2 1,183
- 23 UNI-1.1 1,183
- 24 Grok Imagine Image 1,169
- 25 Recraft V4.1 Utility Pro 1,168
- 26
Cosmos3-Super-Text2Image 1,164
- 27 FLUX.2 [max] 1,161
- 28 Grok Imagine Image Pro 1,161
- 29 FLUX.2 [flex] 1,156
- 30 Reve v1.5 1,154
- 31 FLUX.2 [pro] 1,153
- 32
HunyuanImage 3.0 1,151
- 33
Seedream 4.5 1,149
- 34
Gemini 2.5 Flash Image Preview (nano-banana) 1,149
- 35
Imagen 4 Ultra (generate-001) 1,146
- 36 FLUX.2 [dev] 1,145
- 37
Seedream 5.0 Lite 1,140
- 38
Seedream 4.0 1,139
- 39
Wan 2.6 1,138
- 40 Recraft V4.1 Pro 1,131
- 41
Imagen 4 (generate-001) 1,128
- 42 Krea 2 Medium 1,128
- 43
Qwen-Image (2512) 1,125
- 44
Wan 2.5 Preview 1,117
- 45 Krea 2 Large 1,115
- 46 GPT-Image-1 1,115
- 47 Recraft V4 1,115
- 48
HiDream-O1-Image 1,115
- 49 Krea 2 Turbo 1,114
- 50 GPT-Image-1 mini 1,109
- 51
Wan 2.7 Image Pro 1,103
- 52
Wan 2.7 Image 1,101
- 53 Recraft V4.1 Flash 1,098
- 54
MAI-Image-1 1,092
- 55
Z-Image-Turbo 1,082
- 56
Seedream 3.0 1,082
- 57 FLUX.1 Kontext [max] 1,075
- 58 FLUX.2 [klein] 9B 1,070
- 59
Qwen-Image 1,061
- 60 FLUX.1 Kontext [pro] 1,059
- 61
Imagen 3 (generate-002) 1,057
- 62 Ideogram 3.0 1,049
- 63 Luma Photon 1,036
- 64 P-Image 1,032
- 65 FLUX.2 [klein] 4B 1,030
- 66 Runway Gen-4 1,024
- 67 Recraft V3 1,022
- 68 FLUX1.1 [pro] 1,016
- 69 Ideogram 2.0 1,013
- 70
Lucid Origin 1,013
- 71 GLM-Image 1,012
- 72
Gemini 2.0 Flash Preview Image Generation 976
- 73 FLUX.1 [dev] fp8 969
- 74 DALL-E 3 967
- 75 FLUX.1 Kontext [dev] 942
- 76
Stable Diffusion 3.5 Large 938
- 77
BAGEL 897
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-Image-2.5 Sunburst | 1,427 | Oct 2026 | Official leaderboard | |
| 2 | GPT-Image-2.5 Flare | 1,401 | Oct 2026 | Official leaderboard | |
| 3 | GPT-Image-2 | 1,381 | medium | Oct 2026 | Official leaderboard |
| 4 | | 1,329 | Oct 2026 | Official leaderboard | |
| 5 | | 1,325 | Oct 2026 | Official leaderboard | |
| 6 | Grok Imagine Image 2.0 | 1,317 | Oct 2026 | Official leaderboard | |
| 7 | Reve 2.1 | 1,300 | Oct 2026 | Official leaderboard | |
| 8 | | 1,269 | Oct 2026 | Official leaderboard | |
| 9 | Reve 2.0 | 1,268 | Oct 2026 | Official leaderboard | |
| 10 | | 1,257 | Oct 2026 | Official leaderboard | |
| 11 | | 1,254 | Oct 2026 | Official leaderboard | |
| 12 | | 1,253 | Oct 2026 | Official leaderboard | |
| 13 | | 1,250 | Oct 2026 | Official leaderboard | |
| 14 | | 1,249 | Oct 2026 | Official leaderboard | |
| 15 | | 1,246 | Oct 2026 | Official leaderboard | |
| 16 | GPT-Image-1.5 | 1,236 | high fidelity | Oct 2026 | Official leaderboard |
| 17 | | 1,232 | Oct 2026 | Official leaderboard | |
| 18 | | 1,225 | Oct 2026 | Official leaderboard | |
| 19 | Ideogram 4.0 | 1,200 | Oct 2026 | Official leaderboard | |
| 20 | | 1,191 | Jun 2026 | Official leaderboard | |
| 21 | UNI-1.1 Max | 1,189 | Oct 2026 | Official leaderboard | |
| 22 | | 1,183 | Oct 2026 | Official leaderboard | |
| 23 | UNI-1.1 | 1,183 | Oct 2026 | Official leaderboard | |
| 24 | Grok Imagine Image | 1,169 | Oct 2026 | Official leaderboard | |
| 25 | Recraft V4.1 Utility Pro | 1,168 | Oct 2026 | Official leaderboard | |
| 26 | | 1,164 | Oct 2026 | Official leaderboard | |
| 27 | FLUX.2 [max] | 1,161 | Oct 2026 | Official leaderboard | |
| 28 | Grok Imagine Image Pro | 1,161 | Oct 2026 | Official leaderboard | |
| 29 | FLUX.2 [flex] | 1,156 | Oct 2026 | Official leaderboard | |
| 30 | Reve v1.5 | 1,154 | Oct 2026 | Official leaderboard | |
| 31 | FLUX.2 [pro] | 1,153 | Oct 2026 | Official leaderboard | |
| 32 | | 1,151 | Oct 2026 | Official leaderboard | |
| 33 | | 1,149 | Oct 2026 | Official leaderboard | |
| 34 | | 1,149 | Oct 2026 | Official leaderboard | |
| 35 | | 1,146 | Oct 2026 | Official leaderboard | |
| 36 | FLUX.2 [dev] | 1,145 | Oct 2026 | Official leaderboard | |
| 37 | | 1,140 | Oct 2026 | Official leaderboard | |
| 38 | | 1,139 | Oct 2026 | Official leaderboard | |
| 39 | | 1,138 | Oct 2026 | Official leaderboard | |
| 40 | Recraft V4.1 Pro | 1,131 | Oct 2026 | Official leaderboard | |
| 41 | | 1,128 | Oct 2026 | Official leaderboard | |
| 42 | Krea 2 Medium | 1,128 | Oct 2026 | Official leaderboard | |
| 43 | | 1,125 | Oct 2026 | Official leaderboard | |
| 44 | | 1,117 | Oct 2026 | Official leaderboard | |
| 45 | Krea 2 Large | 1,115 | Oct 2026 | Official leaderboard | |
| 46 | GPT-Image-1 | 1,115 | Oct 2026 | Official leaderboard | |
| 47 | Recraft V4 | 1,115 | Oct 2026 | Official leaderboard | |
| 48 | | 1,115 | Oct 2026 | Official leaderboard | |
| 49 | Krea 2 Turbo | 1,114 | Oct 2026 | Official leaderboard | |
| 50 | GPT-Image-1 mini | 1,109 | Oct 2026 | Official leaderboard | |
| 51 | | 1,103 | Oct 2026 | Official leaderboard | |
| 52 | | 1,101 | Oct 2026 | Official leaderboard | |
| 53 | Recraft V4.1 Flash | 1,098 | Oct 2026 | Official leaderboard | |
| 54 | | 1,092 | Oct 2026 | Official leaderboard | |
| 55 | | 1,082 | Oct 2026 | Official leaderboard | |
| 56 | | 1,082 | Oct 2026 | Official leaderboard | |
| 57 | FLUX.1 Kontext [max] | 1,075 | Oct 2026 | Official leaderboard | |
| 58 | FLUX.2 [klein] 9B | 1,070 | Oct 2026 | Official leaderboard | |
| 59 | | 1,061 | Oct 2026 | Official leaderboard | |
| 60 | FLUX.1 Kontext [pro] | 1,059 | Oct 2026 | Official leaderboard | |
| 61 | | 1,057 | Oct 2026 | Official leaderboard | |
| 62 | Ideogram 3.0 | 1,049 | Oct 2026 | Official leaderboard | |
| 63 | Luma Photon | 1,036 | Oct 2026 | Official leaderboard | |
| 64 | P-Image | 1,032 | Oct 2026 | Official leaderboard | |
| 65 | FLUX.2 [klein] 4B | 1,030 | Oct 2026 | Official leaderboard | |
| 66 | Runway Gen-4 | 1,024 | Oct 2026 | Official leaderboard | |
| 67 | Recraft V3 | 1,022 | Oct 2026 | Official leaderboard | |
| 68 | FLUX1.1 [pro] | 1,016 | Oct 2026 | Official leaderboard | |
| 69 | Ideogram 2.0 | 1,013 | Oct 2026 | Official leaderboard | |
| 70 | | 1,013 | Oct 2026 | Official leaderboard | |
| 71 | GLM-Image | 1,012 | Oct 2026 | Official leaderboard | |
| 72 | | 976 | Oct 2026 | Official leaderboard | |
| 73 | FLUX.1 [dev] | 969 | fp8 | Oct 2026 | Official leaderboard |
| 74 | DALL-E 3 | 967 | Oct 2026 | Official leaderboard | |
| 75 | FLUX.1 Kontext [dev] | 942 | Oct 2026 | Official leaderboard | |
| 76 | | 938 | Oct 2026 | Official leaderboard | |
| 77 | | 897 | Oct 2026 | Official leaderboard |
3 models have held number one since its release in Dec 2024
Lab claims count from the day the card came out, independent runs from the model’s release.
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
No lab model card reports it yet
We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark.
More 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. Kimi K3.1 Pro 84.1%
- 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%
- 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%
- DocVQA In use Models answer 50,000 questions about 12,000 scanned document images by reading their text and layout. Scoring is ANLS, a string similarity to reference answers.
Seed1.5-VL 96.9%