Black Forest Labs

FLUX.2 [klein] 9B KV is a KV-cache optimized variant of the Klein 9B model that caches reference image key-value pairs after the first denoising step, skipping redundant computation on subsequent steps. This delivers up to 2.5x faster inference for multi-reference editing tasks while retaining all capabilities of the standard Klein 9B, including sub-second text-to-image and advanced editing in 4 steps.

FLUX.2 [klein] 9B is a 4-step distilled image generation and editing model designed for sub-second inference without sacrificing visual quality. It unifies text-to-image and advanced editing workflows in a single model, making it suitable for interactive applications, real-time previews, and latency-critical production use.

FLUX.2 [klein] 9B Base is the undistilled foundation model of the Klein family, offering full model capacity for image generation and editing. It is optimized for fine-tuning, customization, and post-training workflows where flexibility, control, and maximum training signal are required.

FLUX.2 [klein] 4B Base is a compact undistilled image generation and editing model with an exceptional quality-to-size ratio. It is well suited for local deployment, efficient fine-tuning, and custom pipelines that require flexibility on limited hardware.

FLUX.2 [klein] 4B is a 4-step distilled image generation and editing model optimized for ultra-low latency inference. It delivers near real-time performance with strong visual quality, enabling interactive workflows and responsive production systems on more constrained hardware.

FLUX.2 [max] is a high-precision text to image and image editing model from Black Forest Labs that generates visuals grounded in real-time information via live web search. It delivers maximum prompt adherence with multi-reference editing and state-of-the-art consistency across identities, objects, and details.

FLUX.2 [pro] is a flow-matching latent transformer for precise text-to-image synthesis and reference-guided editing. It supports multi image references, 4MP outputs, and Mistral-based text conditioning for controllable composition and robust iterative edits that preserve structure.

FLUX.2 dev is an open weight text to image and image editing model from Black Forest Labs. It targets developers who need precise control over prompts, references, and iteration. Use it for non commercial research, workflow prototyping, and multi conditioning image pipelines.

FLUX.2 [flex] is a configurable text to image and image editing model built for precise text placement and stable layouts. It exposes sampling and guidance controls and supports up to ten reference images for consistent characters or products across complex compositions.

FLUX.1 Krea [dev] is an open‑weight text‑to‑image model from Black Forest Labs and Krea AI. It targets opinionated aesthetics and realistic photography. Developers can drop it into FLUX.1 dev workflows to build custom generators that avoid the typical AI look.

FLUX.1 Kontext [max] is a high quality text to image model for production workflows. It focuses on prompt accuracy, sharp local edits, and premium typography rendering. Use it for detailed visual design, branded visuals, and consistent character safe image generation.

FLUX.1 Kontext [pro] combines fast text to image generation with precise image editing. It supports reference images, local region edits, and full scene changes while preserving style and character identity. Ideal for iterative workflows in design, product visuals, and storytelling pipelines.

FLUX.1 Kontext [dev] is an open weights image editing model tuned for fast iterative development. It supports local edits and full scene changes from prompts. Use it for style transfer, background swaps, object edits, and character consistent variations in your pipelines.

FLUX.1 Expand [pro] is an outpainting model that extends images beyond their original frame while preserving structure, lighting, and style. It supports controlled expansion from real or generated inputs and integrates into image editing or generative workflows that need precise, coherent borders.

FLUX.1 Fill [dev] is an open image editing model for inpainting and outpainting with text guidance and masks. It can replace objects, extend scenes, and adjust regions while preserving context. Ideal for pipelines that need controllable edits on real or generated images.

FLUX.1.1 [pro] Ultra is a high resolution text to image model from Black Forest Labs. It generates images up to 4 megapixels in about 10 seconds. Ultra mode targets sharp outputs. Raw mode targets natural photographic style. Built for API integration in real products.

FLUX.1 Canny [dev] is a 12B parameter rectified flow transformer for image generation. It takes a text prompt and an input image. It extracts canny edges as structural guidance. It then generates new images that follow the original composition while applying the prompt.

FLUX.1 Fill Pro provides advanced inpainting and outpainting for real and generated images. Supply an input image, mask, and text prompt. The model fills or extends regions with seamless content that matches context and style. Ideal for edits, layout fixes, and content-aware expansion.

FLUX.1.1 Pro is a flagship text to image model from Black Forest Labs. It improves on FLUX.1 with sharper detail, stronger prompt adherence, and faster sampling. Ideal for production image pipelines, product visuals, and creative tools that require consistent high quality output.

FLUX.1 [schnell] is an open source text to image model from Black Forest Labs. It uses 4 step distillation for very fast generation with strong visual quality. Ideal for local deployment, rapid prototyping, batch image production, and integration into custom creative pipelines.

FLUX.1 [dev] is a 12B parameter text to image model from Black Forest Labs. It targets high fidelity visual generation for research and non commercial use. Developers can build image apps that need strong prompt following and fine visual detail at high resolution.

FLUX.1 Depth [dev] uses input depth maps as structural guidance for new image generation. It preserves scene geometry and layout while allowing creative control over style and details. Ideal for developers who need stable composition in controlled image synthesis.