News · Black Forest Labs · 2 min read
Event date: · Published:
Black Forest Labs releases FLUX 3 Image Klein, a 5.6B open-weights image model
On October 15, 2026, Black Forest Labs released FLUX 3 Image Klein, an open-weights model for text-to-image generation, editing and multi-reference generation that runs on RTX GPUs with 8GB of VRAM in INT8.
Archive entry: this covers an event from October 15, 2026, and we published it on November 8, 2026. The details come from sources dated at the time and may have changed since.
On October 15, 2026, Black Forest Labs released FLUX 3 Image Klein, an open-weights image model with 5.6B parameters, and described it in its launch post on the BFL blog.
What Black Forest Labs released
- It supports text-to-image generation, image editing and multi-reference generation, with up to ten reference images.
- It shares FLUX 3 Image's 4K output, pixel-level editing and bounding-box controls for placing elements or limiting edits to a region.
- Checkpoints include a base model for fine-tuning and a distilled model for fast generation, plus FP8, INT8 and NVFP4 builds made with NVIDIA.
- The INT8 build runs on NVIDIA GeForce RTX GPUs with as little as 8GB of VRAM through ComfyUI or Diffusers.
- Weights were published on Hugging Face as
black-forest-labs/flux-3-image-klein-baseandblack-forest-labs/flux-3-image-klein.
What it means for people who build with AI
The FLUX Kommunity License allows free commercial use for individuals and companies with annual revenue under $5M, while larger companies use self-hosted licensing or the BFL API. Together with an INT8 build that fits a GPU with 8GB of VRAM, this gave indie developers and small companies a commercially usable open model for image generation and editing on consumer GPUs.
API, tooling and internal evaluations
- The BFL API model ID is
flux-3-image-klein, with pricing starting at $0.009 per 1K image. - Launch-day support covered ComfyUI, Hugging Face Diffusers, TensorRT-LLM, SGLang Diffusion and vLLM-Omni, and LoRA training was possible through Ostris AI Toolkit and NVIDIA NeMo AutoModel.
- In BFL's internal evaluations the model scored +91 Elo on text-to-image and +70 Elo on editing against
FLUX.2 [klein] 9B, with median API latency of 4.54 s for text-to-image and 7.18 s for editing at 1 MP.
Sources
- FLUX 3 Image Klein: open weights, 5.6B parameters | Black Forest Labs (Black Forest Labs, October 15, 2026, official)
Written with AI assistance, strictly from the sources listed above. Before it went live, automated checks verified its dates, sources and every language edition.