FULL STORY
Community Races to Speed Up Qwen-Image-2.1
Within a week of Qwen-Image-2.1's release, the community shipped multiple few-step LoRAs built on DMD distillation, speeding up image generation by up to 6.3x.
2026-09-25 ~ 2026-09-26 · 2 episodes · 6 posts
Episode 1 · Pruna Open-Sources Few-Step LoRAs Speeding Up Qwen-Image-2.1 Up to 6.3x (2026-09-25, 4 posts)
PrunaAI open-sourced Pruna-Qwen-Image-2.1, a set of DMD-distilled few-step LoRA adapters that cut Qwen-Image-2.1 image generation and editing from 40 steps down to 5, delivering up to 6.3x speedups, with an online Space for browser testing.
- PrunaAI releases few-step LoRAs for Qwen-Image-2.1, ~5x faster image generation — sdiazlor · 2026-09-25
- Pruna open-sources few-step LoRA adapters making Qwen-Image-2.1 up to 6.3x faster — huggingface · 2026-09-25
- Pruna's few-step LoRAs make Qwen-Image-2.1 up to 6.3x faster in 5 or 8 steps — _akhaliq · 2026-09-25
- Pruna launches browser-based few-step LoRA adapters for Qwen-Image-2.1, 5-step generation — _akhaliq · 2026-09-26
Episode 2 · Viggle's Distilled Qwen-Image-2.1 Generates 5x Faster in 6 Steps (2026-09-25, 2 posts)
Within a week of Qwen-Image-2.1's release, the community shipped three few-step acceleration LoRAs; Viggle's DMD-distilled turbo version cuts generation from 40 steps to 6, about 5x faster.
- Three few-step LoRAs for Qwen Image 2.1 ship within a week of release — linoy_tsaban · 2026-09-25
- Viggle ships 6-step distilled Qwen-Image-2.1 turbo, ~5x faster than 40-step base — _akhaliq · 2026-09-26