Skip to content
IFAI

Model rankings

1133 models · 5 boards · 2 sources

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

Text-to-image arena · Anonymous head-to-head votes · Updated April 19, 2026

Best images
GPT Image 1.5 (high)OpenAI1,275Score
Best open model
FLUX.2 [dev] TurboFal1,167Score
Best from China
Seedream 4.0ByteDance Seed1,202Score
Top three by Score
3Riverflow 2.0Sourceful1,262
Top three open models
1FLUX.2 [dev] TurboFal · #15 overall1,167
2Qwen Image Max 2512Alibaba · #16 overall1,161
3FLUX.2 [dev]Black Forest Labs · #18 overall1,158
This week's changes
claude-opus-4-7Coding · #1New
Gemini 3.1 Flash TTSSpeech · #2New
ImagineArt 2.0Image · #11New
AsyncFlow V2, asyncSpeech · #32New
Lightning v3.1Speech · #46New
RankChangeModelScore95% interval
1—Qwen Image Max 2512OpenAlibaba1,161±10
2—HunyuanImage 3.0 (Fal)OpenTencent1,133±9
3—Z-Image TurboOpenAlibaba1,113±9
4—HunyuanImage 3.0 Instruct (Fal)OpenTencent1,106±10
5—SRPOOpenTencent1,079±9
6—HunyuanImage 2.1OpenTencent1,078±9
7▲ 1HiDream-I1-DevOpenHiDream1,063±10
8▲ 1GLM-ImageOpenZ AI1,059±9
9▼ 2Qwen ImageOpenAlibaba1,056±9
10—HiDream-I1-FastOpenHiDream1,055±9
11—Infinity 8BOpenByteDance1,048±9
12—LongCat ImageOpenMeituan1,045±10
13▲ 1SDXL LightningOpenByteDance907±9
14▼ 1BagelOpenByteDance906±11
15—Janus ProOpenDeepSeek707±13

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 “—”.