← All updates

Qwen Ships Faster Qwen-Image Turbo

The accelerated 7B checkpoint that cuts generation down to eight denoising steps is downloadable now

Hi!

DIT Model Radar tracks new AI models daily: real models or confirmed codenames only, no tools, no hype.

Qwen's Turbo checkpoint of Qwen-Image-2.1 is live on Hugging Face, a 7B text-to-image and image-editing model that generates in just eight denoising steps. It loads through the QwenImage21Pipeline in Diffusers, ships with its own sampling schedule, and defaults to CFG=1 with prefix KV caching.

What changed: A new accelerated checkpoint, Qwen-Image-2.1-Turbo, is published under the official Qwen organisation on Hugging Face. It keeps the same 7B visual generation architecture as Qwen-Image-2.1 but runs in 8 denoising steps with the recommended sampling schedule already included.

  • Open weights: free to download on Hugging Face under the qwen-research license
  • Recommended sampling schedule included; CFG=1 by default with prefix KV caching

It gives local and hosted users a faster route to Qwen-Image-2.1 output without changing their existing pipeline.

Read more →

[ $1 LAUNCH SPECIAL ]
Qwen-Image-2.1-Turbo on DIT
15× the images. Just $1.
≈333 images.*
Unlocks at launch. *Estimated at typical image pricing; launch price TBD.

Top AI models. Lower costs.

Access GPT, Claude, Gemini, and more through one OpenAI-compatible API—and save up to 70% on AI API costs. Create your DIT account, get your API key, and start building.