SGLang-Diffusion Adds Day-0 Support for Qwen-Image-2.1

On September 21, 2026, SGLang-Diffusion announced day-0 support for Alibaba Qwen's Qwen-Image-2.1 model. The diffusion model serving framework now supports text-to-image generation, multi-image editing, and transparent RGBA output. According to benchmarks shared by the team, a single RTX 4090 24GB GPU with CPU offload can run at native precision, generating a 1024×1024 image in 18.7 seconds and performing image editing in 21.7 seconds, with peak GPU memory usage of 22.7 GiB. On an RTX PRO 6000 96GB GPU, generation takes 8.0 seconds.