Qwen Image 2

• 🆕 Alibaba's Qwen image-generation model, released March 2026.
• 🎯 Focused on text rendering inside images and compositional control.
• 👁️ Successor to the original Qwen Image foundation model.
• 🌐 Offered through Alibaba Cloud Model Studio for text-to-image.
• 🏢 Ships alongside a higher-fidelity Pro variant.
• 🔧 Reference and instruction-based editing via companion Edit models.

Alibaba

Alibaba Group is a Chinese multinational technology company founded in 1999 and headquartered in Hangzhou, Zhejiang. Originally built around e-commerce and cloud computing, Alibaba has become one of the most prolific contributors to open-weight AI research,…

Qwen Image 2 is Alibaba's text-to-image generation model, released in March 2026 as the successor to Qwen Image, the lab's earlier image foundation model from January 2025. According to its catalog description, the model's stated design focus is text rendering inside images together with strong compositional control, continuing the emphasis on legible typography and layout that defined the original release.

Alibaba distributes Qwen text-to-image models through its Model Studio platform, which documents generating images directly from prompts. As with its predecessor, the design goal centers on rendering text and structured visual content reliably within a single generation pass, rather than treating text as an afterthought.

The 2.0 release is organized as a small family. Alongside the standard model, Qwen Image 2 Pro is offered as a higher-tier variant, while dedicated editing models, Qwen Image 2 Edit and Qwen Image 2 Pro Edit, extend the same lineage to instruction-based and reference editing workflows.

For factual context, the details here are drawn from Alibaba's Model Studio documentation and the catalog entry. Independent benchmark comparisons against other models were not available among trustworthy evaluators at the time of writing, so no performance rankings are stated.