Skip to content
IFAI

Model rankings

1396 models · 5 boards · 2 sources

Artificial AnalysisText-to-image arenaArenaText-to-image arena

Text-to-image arena · Anonymous head-to-head votes · Published August 7, 2026

Best images
gpt-image-2 (medium)OpenAI1,380Score
Best open model
ideogram-4.0-qualityIdeogram1,206Score
Best from China
qwen-image-3.0-proAlibaba1,258Score
Top three by Score
2reve-2.1Reve1,302
3muse-imageMeta1,283
Top three open models
1ideogram-4.0-qualityIdeogram · #14 overall1,206
2Cosmos3-Super-Text2Image (Agentic)NVIDIA · #18 overall1,181
3hunyuan-image-3.0Tencent · #27 overall1,151
This week's changes
muse-spark-1.2 (xHigh)Coding · #13New
gemini-3.5-flash-highCoding · #25New
Inkling SmallIntelligence · #36New
Celeris-1Intelligence · #221New
RankChangeModelScore95% intervalVotes
1—gpt-image-2 (medium)ClosedOpenAI1,380±569,194
2—reve-2.1ClosedReve1,302±87,598
3—muse-imageClosedMeta1,283±714,510
4—reve-2.0ClosedReve1,270±614,650
5—gemini-3.1-flash-image (nano-banana-2) [web-search]ClosedGoogle1,263±527,910
6Newqwen-image-3.0-proClosedAlibaba1,258±103,735
7▲ 4seedream-5.0-proClosedByteDance1,257±526,510
8▼ 2mai-image-2.5ClosedMicrosoft AI1,256±544,940
9▼ 2gemini-3.1-flash-lite-image (nano-banana-2-lite)ClosedGoogle1,251±616,419
10▼ 2gemini-3-pro-image-2k (nano-banana-pro)ClosedGoogle1,246±3138,756
11▼ 2gpt-image-1.5-high-fidelityClosedOpenAI1,239±3142,427
12▼ 2gemini-3-pro-image-preview (nano-banana-pro)ClosedGoogle1,232±582,835
13▼ 1grok-imagine-image-qualityClosedxAI1,228±448,786
14▼ 1ideogram-4.0-qualityOpenIdeogram1,206±527,891
15▼ 1qwen-image-2.0-pro-2026-06-22ClosedAlibaba1,191±612,026
16▼ 1uni-1.1-maxClosedLuma AI1,188±613,491
17▼ 1mai-image-2ClosedMicrosoft AI1,182±549,207
18▲ 38Cosmos3-Super-Text2Image (Agentic)OpenNVIDIA · 1 more variant1,181±102,957
19▼ 2uni-1.1ClosedLuma AI1,180±526,927
20▼ 2grok-imagine-imageClosedxAI1,172±3217,461
21▼ 2recraft-v4.1-utility-proClosedRecraft1,169±112,519
22▼ 2flux-2-maxClosedBlack Forest Labs1,162±4117,464
23▼ 2grok-imagine-image-proClosedxAI1,161±493,711
24▼ 2flux-2-flexClosedBlack Forest Labs1,156±3149,402
25▼ 2flux-2-proClosedBlack Forest Labs1,155±3175,196
26▼ 2reve-v1.5ClosedReve1,154±436,799
27▼ 1hunyuan-image-3.0OpenTencent1,151±3173,366
28▼ 3gemini-2.5-flash-image-preview (nano-banana)ClosedGoogle1,150±3822,940
29▼ 2imagen-ultra-4.0-generate-001ClosedGoogle1,148±4389,010
30▼ 1seedream-4.5ClosedByteDance1,146±3234,737
31▼ 3flux-2-devOpenBlack Forest Labs1,146±466,740
32▼ 2seedream-4-2kClosedByteDance1,140±712,639
33▼ 1seedream-5.0-liteClosedByteDance1,136±496,040
34▼ 3wan2.6-t2iClosedAlibaba1,136±3182,381
35▼ 2recraft-v4.1-proClosedRecraft1,130±112,682
36▼ 2imagen-4.0-generate-001ClosedGoogle1,129±3532,862
37▼ 2qwen-image-2512OpenAlibaba · 1 more variant1,126±491,329
38▼ 2krea-2-mediumClosedKrea1,122±526,584
39▼ 2hidream-o1-imageOpenHiDream1,118±434,179
40▼ 1wan2.5-t2i-previewClosedAlibaba1,117±3230,234
41▼ 3seedream-4-falClosedByteDance1,116±711,911
42▼ 2gpt-image-1ClosedOpenAI1,115±3265,373
43—recraft-v4ClosedRecraft1,113±499,721
44▼ 2seedream-4-high-res-falClosedByteDance1,113±3175,579
45▼ 4krea-2-turboOpenKrea1,110±618,650
46▼ 2gpt-image-1-miniClosedOpenAI1,109±3164,065
47▼ 1krea-2-largeClosedKrea1,105±526,191
48▼ 3wan2.7-image-proClosedAlibaba1,103±528,745
49▼ 2wan2.7-imageClosedAlibaba1,100±529,087
50▼ 2mai-image-1ClosedMicrosoft AI1,093±494,700

Intelligence over time

Each maker's best model at the time on the Artificial Analysis Intelligence Index, by the month it was released.

Gained most · Anthropic +57.2Highest now · Anthropic 64.6Gap between highest and lowest 18.1 ptsTime range Nov 2025 – Nov 2026
64.6Anthropic60.0Google56.9OpenAI50.1Meta46.5Kimi

Arena data comes from the leaderboard dataset it publishes on Hugging Face (CC BY 4.0). Only each board's overall leaderboard is used; scores and ranks are unchanged. dataset · CC BY 4.0

Artificial Analysis data comes from its API (artificialanalysis.ai). Scores are unchanged; ranks follow the scores. artificialanalysis.ai

Open or closed follows the license Arena gives; prices and Artificial Analysis intelligence show only where model names match, otherwise “—”.