Qwen Image

🚀 Meet Qwen-Image — a 20B MMDiT model for next-gen text-to-image generation. Especially strong at creating stunning graphic posters with native text. Now open-source.

🔍 Key Highlights:
🔹 SOTA text rendering — rivals GPT-4o in English, best-in-class for Chinese
🔹 In-pixel… pic.twitter.com/zT9CFLzWkV

— Qwen (@Alibaba_Qwen) August 4, 2025

All of these images were made w/ Qwen-Image on @replicate, big W @Alibaba_Qwen https://t.co/RYJEFzTomz pic.twitter.com/s1rKUaQSt9

— Sakib (@zsakib_) August 5, 2025

KAMUI CODE 解説AI動画。

Qwen imageはVivid（鮮やか）レベルが結構効くらしいので、やってみた。

いい感じ！

文字はomnihumanでi2vしたら潰れたwww#kamuicode pic.twitter.com/coc58gfqPH

— 神威/KAMUI (@kamui_qai) August 5, 2025

今日もお疲れさまでした
日本の最高気温が更新されましたね☀️https://t.co/MoNYvtLrrU
手書き文字もいけそうだが漢字が一部中国語に Qwen Image RTX3060にて #Qwen pic.twitter.com/6fuM9QYwdl

— 国崎犀考 (@saikou_kunisaki) August 5, 2025

QWEN Image is really good at text rendering, but sometimes it writes its own name as “QWON.” https://t.co/98pF9tfqZM pic.twitter.com/giBfVJ8tfT

— İlker (@ilkerigz) August 4, 2025

Qwen-Image is now the leading open weights Text to Image model.

It's text rendering is 👌

---
Prompt: Create a motivational poster in the style of Saul Bass - bold forms, strong hand-drawn typography. The text says: "Imagine the future as you wish it to be" second part "because… pic.twitter.com/OHUQGMQAj8

— frasernz (@MediaCentralDev) August 4, 2025

Qwen team is about to blow you away, ONCE AGAIN!

Imagine if you formulated image generation as a GRPO problem... (and what if it's better than GPT-Image?)

Here is a Cyberpunk-inspired gallery:👇 pic.twitter.com/olIRaghRe5

— Casper Hansen (@casper_hansen_) August 4, 2025