Optimal Settings: 6 inference steps, and a cfg scale of 1.0.
FLUX.2 [klein]: Real-time Image Generation with Uncompromised Quality
FLUX.2 [klein] 9B is a 9 billion parameter rectified flow transformer designed to generate images from text descriptions. This flagship small model pushes the Pareto frontier for quality versus latency in text-to-image generation. It matches or even surpasses the performance of models five times its size, achieving this in under half a second.
Lightning version
FLUX.2 [klein] 9B is built upon a 9B flow model and integrates an 8B Qwen3 text embedder, further optimized through step-distillation to just 4 inference steps.
Key Features of FLUX.2 [klein]
Sub-second Image Generation: A distilled model designed for rapid image creation without compromising on quality.
Text-to-Image Capabilities: Generates images directly from textual prompts.
Creative Exploration: Offers strong prompt adherence and diverse output, making it suitable for creative workflows.
Experience the speed and quality of FLUX.2 [klein] yourself. Start generating images in real-time now!
