Qwen

A 20-billion-parameter multimodal diffusion transformer model excelling in complex multilingual text rendering, precise image generation, and advanced editing like style transfer, object manipulation, and pose adjustment with strong semantic and visual fidelity.