Release2026-10-28
Unsloth released a GGUF-quantized version of FLUX.2-klein-9B. It uses Unsloth Dynamic 2.0 for state-of-the-art performance. The model supports text-to-image and image-to-image editing in one architecture. It delivers sub-second inference on consumer GPUs. It fits in 29GB VRAM. It is available for non-commercial use only. It is built on ComfyUI-GGUF tooling. It is accessible via Hugging Face, ComfyUI, Diffusers and the BFL API.
What changed
FLUX.2-klein-9B was available only in its original, larger format.
What it unlocks
Running a 9B-parameter image model on consumer GPUs with sub-second generation.
- 9B parameters
- 4 inference steps
- 29GB VRAM
- 409,688 downloads last month
What you need to act on it
- a GPU with at least 29GB VRAM
- non-commercial use only
Sources