# Black Forest Labs FLUX.3 Image Klein arrives in Diffusers

> **Key takeaway:** The new FLUX.3 Image Klein model supports text-to-image, multi-reference image-to-image, and bounding-box layout control.

**Published:** 2026-10-15T00:00:00+00:00 
**Source:** Hugging Face Blog & Announcements 
**Category:** ai-open-models 
**Canonical URL:** https://fosswire.org/news/black-forest-labs-flux3-image-klein-arrives-in-diffusers.html 

## Summary
Black Forest Labs has released FLUX.3 Image Klein, an open image generation model now compatible with the Diffusers library. The model introduces layout control via bounding boxes, allowing users to specify object placement or movement within a scene. It is available in both a base version and a distilled version optimized for fast inference.

## Technical capabilities of FLUX.3
- **Layout Control**: Uses JSON-formatted bounding boxes [y_min, x_min, y_max, x_max] on a 0–1000 scale for precise object positioning.
- **Multi-reference Image-to-Image**: Supports multiple input images referenced via tags like `<ref_image_0>` in the prompt.
- **Two Model Variants**:
- `flux-3-image-klein`: Distilled for speed (16 steps, no CFG), capable of 1024x1024 generation in ~2 seconds on an H100.
- `flux-3-image-klein-base`: Standard model for fine-tuning (50 steps, with CFG).
- **Quantization Support**: Supports NVFP4 and FP8 weights for reduced VRAM usage on smaller GPUs.
- **Modular Pipeline**: Implemented as a Modular Diffusers pipeline for easier component swapping.

## Why it matters
The integration into Diffusers and the availability of LoRA training scripts provide developers with high-control tools for open-source image generation workflows.