blackforestlabs/flux-2/dev/edit

Transform and refine images with sharp detail, precise lighting, and text accuracy using Flux 2 Dev image-to-image, a high-resolution, open-weight model for professional creative workflows.

The URLs of the images for editing. Maximum of 3 image URLs; if more are provided, only the first 3 will be used.
Guidance Scale is a measure of how close you want the model to stick to your prompt when generating the image.
The number of inference steps to perform during image generation.
Specifies the size of the image to generate. Choose a preset aspect ratio or use Custom to specify width and height between 512 and 2048 pixels.
The number of images to generate.
If set to true, the prompt will be expanded for better results.
Specifies the output image format.

Introduction To Flux 2 Dev Image Editor

Flux 2 Dev image-to-image is the latest open-weight model in the FLUX.2 family, created by Black Forest Labs with an impressive 32-billion parameter architecture. It blends a rectified flow transformer and vision-language model backbone to deliver highly detailed visuals, sharper textures, and consistent lighting control. You can run Flux 2 Dev locally or through APIs, giving you complete flexibility to build and deploy creative tools. Compared to its predecessors, this upgraded version offers superior text rendering, high-resolution editing up to four megapixels, and the ability to manage up to ten reference images for unmatched consistency and realism. Flux 2 Dev image-to-image lets you transform or refine pictures with precision, designed for creators, artists, and visual teams who need compelling, brand-consistent results. You can modify photos while preserving detail, generate photorealistic outputs, and streamline your entire creative workflow with ease using this professional-grade image generation tool.

Visual Examples Created With Flux 2 Dev

Flux 2 Dev on X: Updates and Insights

What makes Flux 2 Dev stand out

Flux 2 Dev is a high-fidelity image-to-image editor built to preserve structure while delivering sharp detail, precise lighting, and dependable text rendering. Instead of full-frame regeneration, Flux 2 Dev applies targeted, content-aware changes that keep geometry, materials, and perspective intact. The open-weight design of Flux 2 Dev supports rigorous professional workflows, enabling reproducible results with seeds, stable guidance, and multi-image conditioning for style or reference transfer. Optimized for high-resolution output, Flux 2 Dev maintains realism under substantial restyling, from subtle retouching to bold aesthetic shifts. With efficient inference and configurable steps, Flux 2 Dev balances speed and quality for production use. Key capabilities:

  • Flux 2 Dev preserves pose, layout, depth cues, and material response to avoid warping and unwanted resynthesis.
  • Lighting coherence across edits, with shadows, reflections, and color temperature that remain physically plausible.
  • Reliable typography replacement with scene-aware placement, perspective consistency, and stable kerning.
  • Multi-image conditioning supports up to 3 references for style transfer, palette alignment, and motif continuity.
  • High-resolution fidelity with crisp edges, fine textures, and clean gradients across common aspect ratios.
  • Tunable behavior in Flux 2 Dev via guidance, steps, seeds, and acceleration for predictable iteration.

Prompting guide for Flux 2 Dev

With Flux 2 Dev, start with a high-quality base image and a clear instruction that states what to change and what to preserve. Provide up to three image URLs to guide style or content, then describe the edit target, scope, and desired lighting or typography. In Flux 2 Dev, use guidance_scale to control adherence, num_inference_steps to trade speed for detail, and image_size for aspect and resolution. Set seed for reproducibility, enable Magic Prompt only when you want expanded phrasing, and choose output_format. For fast previews, keep acceleration at regular; for maximum quality, raise steps moderately. Flux 2 Dev respects precise constraints, so phrase exclusions explicitly. Examples:

  • In Flux 2 Dev, preserve subject and composition; replace background with a minimalist studio wall and soft key light.
  • Add a wooden bench to the right of the subject; background only; maintain natural shadows and perspective.
  • Replace storefront sign text with "ATELIER", white sans-serif, perspective matched; no other changes.
  • Use Flux 2 Dev to transfer color palette from reference 2; keep clothing texture and folds unchanged.
  • Global restyle to watercolor look while keeping edges crisp and typography legible. Pro tips:
  • State hard constraints explicitly, for example "do not alter face", to keep scope narrow.
  • Use spatial language such as "left of subject", "upper-right quadrant", or "background only".
  • Avoid overstuffed adjectives; prefer 3 to 5 strong descriptors that guide Flux 2 Dev predictably.
  • Iterate by adjusting guidance_scale in small steps and nudging the prompt; reuse seed to compare results.
  • Supply high-resolution, on-topic references; crop irrelevant regions before sending to Flux 2 Dev.

Related Playgrounds

Frequently Asked Questions

What is Flux 2 Dev and what does its image-to-image feature do?

Flux 2 Dev is a powerful 32-billion parameter open-weight generative model designed for image creation and editing. Its image-to-image feature allows users to modify existing pictures using prompts while preserving key parts of the composition.

How is Flux 2 Dev different from previous versions like FLUX.1 in terms of image-to-image performance?

Flux 2 Dev provides sharper textures, more realistic lighting, and finer detail fidelity compared to FLUX.1, making its image-to-image transformations noticeably higher in quality and accuracy.

Is Flux 2 Dev free to use, and how does pricing work for image-to-image generation?

Flux 2 Dev offers a free trial tier with complimentary credits for new users. After that, you can use a credit-based pricing model where each image-to-image or text-to-image task consumes credits proportional to the image size.

Who are the main users of Flux 2 Dev, and how might they benefit from its image-to-image capabilities?

Flux 2 Dev is ideal for designers, advertisers, game studios, and product photographers who require consistent visuals. The image-to-image feature helps them efficiently update or blend styles while keeping brand or visual coherence.

On what platforms can I access Flux 2 Dev and its image-to-image tools?

You can access Flux 2 Dev through the Runcomfy AI Playground on desktop or mobile browsers. The image-to-image editing tools work directly in the browser, with local or API usage options available for developers.

What technical inputs and outputs does Flux 2 Dev support for image-to-image creation?

Flux 2 Dev accepts existing images and textual prompts as inputs and can output edited visuals up to about 4 megapixels. The image-to-image process maintains structure while adding new elements or styles.

Are there any limitations when using Flux 2 Dev for image-to-image editing?

Flux 2 Dev performs best up to around 4 megapixels and may require optimization for extremely complex multi-reference setups. However, the model’s image-to-image guidance is still among the most stable available.

Can I run Flux 2 Dev locally for image-to-image workflows?

Yes, Flux 2 Dev can be run locally using its open weights or accessed remotely through APIs, making image-to-image experimentation flexible for both individual creators and enterprise users.

How does Flux 2 Dev maintain consistency when performing image-to-image tasks with multiple references?

Flux 2 Dev allows up to ten reference images per prompt. This feature ensures character or style consistency across related outputs, giving image-to-image edits a coherent and unified aesthetic.