Qwen Image is Alibaba's open-weight image foundation model: a 20B multimodal diffusion transformer released under Apache 2.0. It renders readable multi-line text inside images in English, Chinese, Japanese, and Korean, which is the one task most diffusion models still get wrong.
In Phygital+ it runs as a single node for both generation and editing. Prompt it from scratch, or connect up to three reference images to guide style, identity, or layout, with camera angle presets and LoRA support built into the same node.
Posters, banners, packaging, and any layout that needs legible text inside the image
Design and marketing teams working across English, Chinese, Japanese, and Korean
Legible multilingual typography, reference-guided editing, and LoRA control in one node.
Qwen Image was trained for complex text rendering: multi-line layouts, paragraph-level copy, and correct spelling in both alphabetic and logographic scripts. Posters, packaging mockups, and ad banners come out with headlines you can use, instead of gibberish you have to patch in Photoshop.
English, Chinese, Japanese, and Korean all render with font detail and layout coherence preserved. For teams shipping the same campaign across several markets, that removes a separate design pass for every visual.
Connect up to three reference images and describe only what should change. The model holds semantic meaning and visual realism during edits, so identity, style, and untouched areas stay put instead of drifting.
Horizontal rotation in 45° steps, three distance settings, and vertical tilt from low angle to high angle give you repeatable framing without rewriting the prompt every time.
Qwen Image ships under Apache 2.0, one of the most permissive licences among large image models. Commercial use is unrestricted at the model level, which matters if you also plan to fine-tune or self-host.
Qwen Image fits wherever a visual has to carry readable text, survive several rounds of edits, or hold the same character and style across a set.
Add the node, write a prompt or connect references, and generate.
Add a QWEN node to the canvas and describe the result in a text prompt. For editing or style control, connect up to three reference images at up to 1664 px per side.
Pick an aspect ratio preset or set width and height directly, from 512 to 1664 px. Use the Angle Settings group for camera rotation, distance, and vertical tilt, and raise Steps toward 30 when you need more detail.
For edits, describe only what should change so the rest of the image stays as it was. Chain the result into an upscaler, a background remover, or a video node, then export with commercial rights on your plan.
One subscription across 30+ AI models, with no per-tool credit balances or separate signups. Credit cost per generation is shown live in the node before you run it.
500 weekly credits to test the model and see the output quality. No credit card required.
45,000 monthly credits with cheaper per-credit pricing, commercial rights, and video download.
90,000 monthly credits, up to 10 seats, shared workspace, centralized billing, and priority support.
210,000 monthly credits, API and integration support, unlimited seats, and dedicated management.
Full specs for Qwen Image in Phygital+.
How Qwen Image compares with other AI image models in the Phygital+ catalog.
Chain generation, editing, upscaling, and video into one repeatable workflow.
Generate at 1664 px, then push the result through an upscaling node to reach print or hero-banner resolution.
Use a Qwen Image frame as the first frame for a video node and animate a layout you have already approved.
Draft the headline and body copy with a text node, then feed it into the prompt so typography and wording are decided in one pass.
Cut the subject out of a generated frame and drop it into a composition, a mockup, or a banner template.
Everything you need to know about Qwen Image in Phygital+.
Qwen Image is the first image generation foundation model from Alibaba's Qwen team, released August 4, 2025. It is a 20B-parameter MMDiT (multimodal diffusion transformer) published under the Apache 2.0 licence, built specifically for complex text rendering inside images and for instruction-based editing.
Text rendering was a training objective rather than a side effect. The model pairs a Qwen2.5-VL text encoder with the diffusion transformer, so layout and typography are modelled alongside the image itself. On text benchmarks such as LongText-Bench, ChineseWord, and TextCraft it outperforms other leading models, with the widest margin on Chinese.
English, Chinese, Japanese, and Korean render with font detail and layout coherence preserved. The language you prompt in and the language of the text inside the image are set separately, so you can write the brief in one language and request in-image copy in another.
Yes. Connect up to three images to the Start Images input and describe only what should change. Being specific about the edit reduces drift, so parts of the image you did not mention stay as they were.
Width and height run from 512 to 1664 pixels, with presets for 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 5:7, and 7:5. For larger output, chain the result into an upscaling node on the same canvas.
Yes. The node has a My lora field for LoRAs you have trained in Phygital+, with adjustable strength, plus a separate Modifier slot for preset LoRAs such as qwen-fusion.
Usage is credit-based and included in every plan, from Free up to Enterprise. The credit cost per generation is shown in the node before you run it, so there is no separate signup or credit balance for this one model.
Outputs created on any paid plan come with commercial-use rights. Qwen Image itself is published under Apache 2.0, which places no additional restriction on commercial use of the model.
Running it in Phygital+ means no separate signup, no GPU to rent, and no per-model credit balance. It sits on the same canvas as 30+ other AI tools, so generation, editing, upscaling, and video can be chained into one repeatable workflow under a single subscription.
Browse the full catalog of 30+ AI models available in Phygital+.
Join 100+ teams using Phygital+ – every model in one workspace
Try Phygital+ free
