Skip to content
IFAI

Model rankings

1492 models · 5 boards · 2 sources

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

Text-to-image arena · Anonymous head-to-head votes · Published September 8, 2026

Best images
gpt-image-2.5-sunburstOpenAI1,421Score
Best open model
ideogram-4.0-qualityIdeogram1,204Score
Best from China
seedream-5.0-proByteDance1,257Score
Top three by Score
2gpt-image-2.5-flareOpenAI1,399
Top three open models
1ideogram-4.0-qualityIdeogram · #17 overall1,204
2Cosmos3-Super-Text2Image (Agentic)NVIDIA · #24 overall1,167
3hunyuan-image-3.0Tencent · #30 overall1,151
This week's changes
GPT-6 Astra (max)Intelligence · #2New
MAI-Image-2.6Image · #2New
Realtime TTS-2Speech · #2New
Muse Spark 1.3 (max)Intelligence · #5New
RankChangeModelScore95% intervalVotes
1Newgpt-image-2.5-sunburstClosedOpenAI1,421±133,149
2Newgpt-image-2.5-flareClosedOpenAI1,399±132,856
3▼ 2gpt-image-2 (medium)ClosedOpenAI1,381±478,731
4▼ 2mai-image-2.6ClosedMicrosoft AI1,331±711,213
5▼ 2grok-imagine-image-2.0 (low)ClosedxAI1,315±122,681
6▼ 2reve-2.1ClosedReve1,301±87,573
7▼ 2muse-imageClosedMeta1,277±626,045
8▼ 2reve-2.0ClosedReve1,270±614,632
9▼ 2gemini-3.1-flash-image (nano-banana-2) [web-search]ClosedGoogle1,261±541,957
10▼ 2seedream-5.0-proClosedByteDance1,257±462,443
11▼ 2qwen-image-3.0-proClosedAlibaba1,254±710,920
12▼ 2mai-image-2.5ClosedMicrosoft AI1,254±457,250
13▼ 2gemini-3.1-flash-lite-image (nano-banana-2-lite)ClosedGoogle1,250±616,371
14▼ 2gemini-3-pro-image-2k (nano-banana-pro)ClosedGoogle1,246±3152,220
15▼ 2gpt-image-1.5-high-fidelityClosedOpenAI1,239±3151,702
16▼ 2gemini-3-pro-image-preview (nano-banana-pro)ClosedGoogle1,232±582,737
17▼ 2ideogram-4.0-qualityOpenIdeogram1,204±442,169
18▼ 2qwen-image-2.0-pro-2026-06-22ClosedAlibaba1,191±612,005
19▼ 2uni-1.1-maxClosedLuma AI1,188±613,469
20▼ 2mai-image-2ClosedMicrosoft AI1,183±549,198
21▼ 2uni-1.1ClosedLuma AI1,181±533,836
22▼ 2grok-imagine-imageClosedxAI1,171±3236,623
23▼ 2recraft-v4.1-utility-proClosedRecraft1,169±112,516
24▼ 2Cosmos3-Super-Text2Image (Agentic)OpenNVIDIA1,167±87,424
25▼ 2flux-2-maxClosedBlack Forest Labs1,162±4117,339
26▼ 2grok-imagine-image-proClosedxAI1,161±493,639
27▼ 2flux-2-flexClosedBlack Forest Labs1,157±3149,288
28▼ 2flux-2-proClosedBlack Forest Labs1,154±3188,083
29▼ 2reve-v1.5ClosedReve1,154±436,748
30▼ 2Cosmos3-Super-Text2ImageOpenNVIDIA1,154±712,499
31▼ 2hunyuan-image-3.0OpenTencent1,151±3173,099
32▼ 2gemini-2.5-flash-image-preview (nano-banana)ClosedGoogle1,150±2845,086
33▼ 2imagen-ultra-4.0-generate-001ClosedGoogle1,148±4388,103
34▼ 2seedream-4.5ClosedByteDance1,147±3259,145
35▼ 2flux-2-devOpenBlack Forest Labs1,145±473,963
36▼ 2seedream-4-2kClosedByteDance1,140±712,621
37▼ 2seedream-5.0-liteClosedByteDance1,138±3109,182
38▼ 2wan2.6-t2iClosedAlibaba1,137±3201,108
39▼ 2recraft-v4.1-proClosedRecraft1,130±112,680
40▼ 2imagen-4.0-generate-001ClosedGoogle1,129±3531,811
41▼ 2qwen-image-2512OpenAlibaba1,125±492,541
42▼ 2krea-2-mediumClosedKrea1,123±540,205
43▼ 2wan2.5-t2i-previewClosedAlibaba1,118±3248,411
44▼ 2hidream-o1-imageOpenHiDream1,117±448,201
45▼ 2seedream-4-falClosedByteDance1,116±711,870
46▼ 2gpt-image-1ClosedOpenAI1,115±3264,897
47▼ 2recraft-v4ClosedRecraft1,114±3114,295
48▼ 2seedream-4-high-res-falClosedByteDance1,113±3175,361
49▼ 2krea-2-turboOpenKrea1,110±533,237
50▼ 2gpt-image-1-miniClosedOpenAI1,109±3163,843

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 “—”.