Qwen Releases Qwen-Image-2.1-Turbo, an Open-Weight Image Model Running in 8 Denoising Steps
On 2026.10.09, Qwen released Qwen Image 2.1 Turbo, an open weight text to image generation and image editing model in the Qwen family that runs in just 8 denoi…
Published on MyPrivateClaw
Oct 10, 2026, 3:20 AM UTC
Coverage date
Oct 9, 2026
Last updated
Oct 10, 2026, 3:20 AM UTC
News summary
On 2026.10.09, Qwen released Qwen Image 2.1 Turbo, an open weight text to image generation and image editing model in the Qwen family. According to the release record, it runs in just 8 denoising steps by reusing the 7B parameter Qwen Image 2.1 visual generation component. The underlying Qwen Image 2.1 model uses 7B parameters in its visual generation component across 32 Single Stream DiT layers, employing mixed granularity attention and prefix KV cache reuse to balance generation quality, inference efficiency, and versatility. The 7B parameter figure refers to the visual generation component only. Qwen Image 2.1 Turbo is available on Hugging Face and ModelScope under the Qwen Research License, and Qwen Image 2.1 Pro and Turbo APIs are officially live on Alibaba Cloud Model Studio for hosted use. This availability is via the vendor's own platforms; third party mirrors are not verified.…