Geração de imagem com IA Qwen Image 2.0

Qwen Image 2.0 is Alibaba's production image generation model and the strongest option on FlyAIgh for any image that needs readable Chinese text — posters, packaging, signage, social cards in CJK contexts. Where most image models struggle with non-Latin scripts (producing noisy lookalike characters), Qwen Image renders clean Chinese characters that match the prompt, and matches GPT Image 2's quality on English in-image text too.

The family ships two variants. Qwen Image 2.0 is the standard model at 3 credits per image — the same price tier as Nano Banana 2 and Seedream. Qwen Image 2.0 Pro is the higher-fidelity variant at 5 credits per image — finer detail, better photorealism, stronger composition handling for complex prompts.

Both variants are unified — text-to-image and image-to-image (edit) run on the same model id with the platform auto-routing based on whether you supply a reference image. This is unusual on FlyAIgh: GPT Image 2, Nano Banana Pro and WAN 2.7 Pro all use separate model ids for t2i vs i2i, and Qwen's unified approach keeps prompt iteration in a single context. You can supply 1 reference image for editing.

Five aspect ratios are supported (1:1, 16:9, 9:16, 4:3, 3:4). Output sizes match the other tier-3 image models — comparable to Nano Banana 2 at 3 credits and to WAN 2.7 Pro at 5 credits.

Qwen Image is particularly strong at Chinese and English text rendering, brand-asset composition, Asian aesthetics and photorealism. It is weaker than Midjourney v7 for distinctive artistic style and weaker than WAN 2.7 Pro for raw 4K detail. Pick Qwen Image when bilingual text legibility or unified t2i/i2i flow matters most.

On FlyAIgh, Qwen Image runs through Alibaba Cloud Bailian without a separate Alibaba Cloud account, API key or Mainland China registration required.